BEGIN:VCALENDAR VERSION:2.0 PRODID:-//FORMAL SYSTEM ANALYSIS//NONSGML Events//EN CALSCALE:GREGORIAN X-WR-CALNAME:FORMAL SYSTEM ANALYSIS - Events X-ORIGINAL-URL:https://fsa.win.tue.nl/events/event/ X-WR-CALDESC:FORMAL SYSTEM ANALYSIS - Events BEGIN:VEVENT UID:20240312T0847Z-1710233269.6712-EO-1525-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20240312T084245Z LAST-MODIFIED:20240312T084245Z DTSTART;TZID=Europe/Amsterdam:20240314T124500 DTEND;TZID=Europe/Amsterdam:20240314T133000 SUMMARY: Jan Friso Groote: Rewriting with 64-bit digits in mCRL2 DESCRIPTION: Rewriting of numbers is often done with the Peano rewrite rule s\, but this is terribly slow and cumbersome. In mCRL2 the standard built i n numbers are based on a binary encoding\, using constructors @c0 for zero and @cDub for duplication. This is much faster. But when many calculations are done\, the manipulation of these binary ...continue reading X-ALT-DESC;FMTTYPE=text/html:
Rewriting of numbers is often done with th e Peano rewrite rules\, but this is terribly slow and cumbersome. In mCRL2 the standard built in numbers are based on a binary encoding\, using constr uctors @c0 for zero and @cDub for duplication. This is much faster. But whe n many calculations are done\, the manipulation of these binary numbers is still sluggish. The question is whether we can employ machine arithmetic. T his can be done by basing the rewrite system on an arbitrary base b. By tak ing b=2^64 we use the machine 64-bit numbers as digits. This can lead to su bstantial speed ups.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Myrthe Spronck":MAILTO:m.s.c.spronck@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-friso-groote-rewritin g-with-64-bit-digits-in-mcrl2/ END:VEVENT BEGIN:VEVENT UID:20240306T0950Z-1709718647.709-EO-1524-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20240306T094657Z LAST-MODIFIED:20240306T094657Z DTSTART;TZID=Europe/Amsterdam:20240307T124500 DTEND;TZID=Europe/Amsterdam:20240307T133000 SUMMARY: Jan Heemstra: Hardware accelerated intelligent theorem proving DESCRIPTION: We present a connection-based tableaux theorem prover that per forms inferences entirely on the GPU. Benchmarks on the m40 dataset show it performs worse than other provers in terms of proving power. In terms of i nferences per second\, however\, it is on par with or even surpasses state- of-the-art provers on similarly priced and dated hardware\, despite ...cont inue reading X-ALT-DESC;FMTTYPE=text/html:We present a connection-based tableaux the orem prover that performs inferences entirely on the GPU. Benchmarks on the m40 dataset show it performs worse than other provers in terms of proving power. In terms of inferences per second\, however\, it is on par with or e ven surpasses state-of-the-art provers on similarly priced and dated hardwa re\, despite it lacking various important optimizations that are present in the other provers. On top of this prover\, we designed and implemented a h euristic neural network-based system that avoids evaluations during the com putation of inferences\, and instead pre-computes the results of these eval uations.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Myrthe Spronck":MAILTO:m.s.c.spronck@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-heemstra-hardware-acc elerated-intelligent-theorem-proving/ END:VEVENT BEGIN:VEVENT UID:20240227T0931Z-1709026296.2355-EO-1511-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20240227T092653Z LAST-MODIFIED:20240227T092653Z DTSTART;TZID=Europe/Amsterdam:20240229T124500 DTEND;TZID=Europe/Amsterdam:20240229T133000 SUMMARY: Simone Tini: Measuring Robustness in Cyber-Physical Systems under Sensor Attacks DESCRIPTION: We propose a formal framework for quantitative analysis of bou nded sensor attacks on cyber-physical systems\, using the formalism of diff erential dynamic logic. Given a precondition and postcondition of a system\ , we formalize two quantitative safety notions\,quantitative forward and ba ckward safety\, which respectively express (1) how strong the strongest pos tcondition of the system is with respect ...continue reading X-ALT-DESC;FMTTYPE=text/html:When verifying liveness properties on tran sition systems\, it is often necessary to discard spurious counterexamples. This can be done through the application of completeness criteria: assumpt ions on which paths represent realistic executions of the modelled system. To support verification of properties under completeness criteria\, we have developed template modal mu-calculus formulae that can be instantiated to capture a broad range of liveness properties under the following completene ss criteria: progress\, justness\, weak fairness\, strong fairness and hype rfairness. In this talk\, I will present the formulae instantiated for the starvation freedom property of mutual exclusion algorithms\, and show how t he completeness criteria have been represented.
This is joint work wi th Bas Luttik and Tim Willemse
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Myrthe Spronck":MAILTO:m.s.c.spronck@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/myrthe-spronck-completene ss-criteria-in-modal-mu-calculus-formulae/ END:VEVENT BEGIN:VEVENT UID:20240205T0911Z-1707124318.9188-EO-1498-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20240205T091153Z LAST-MODIFIED:20240205T091153Z DTSTART;TZID=Europe/Amsterdam:20240208T124500 DTEND;TZID=Europe/Amsterdam:20240208T133000 SUMMARY: Jeroen Keiren: An Expressive Timed Modal Mu-Calculus for Timed Aut omata DESCRIPTION: In the untimed setting\, it is well-known that the modal mu-ca lculus is more expressive than other modal logics such as LTL\, CTL and CTL *. It can thus be considered a foundational logic for model-checking. In th e timed setting\, the status of similarly foundational logics is less satis factory. There are timed extensions of modal logics\, such ...continue read ing X-ALT-DESC;FMTTYPE=text/html:In the untimed setting\, it is well-known
that the modal mu-calculus is more expressive than other modal logics such
as LTL\, CTL and CTL*. It can thus be considered a foundational logic for m
odel-checking.
In the timed setting\, the status of similarly foundati
onal logics is less satisfactory. There are timed extensions of modal logic
s\, such as TCTL. Yet\, the state of the art of timed mu-calculi is underde
veloped.
In this talk\, I will introduce a timed mu-calculus that is m
ore expressive than existing timed mu-calculi\, and that is the first timed
mu-calculus that is more expressive than TCTL over arbitrary timed automat
a.
This is joint work with Rance Cleaveland and Peter Fontana.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Myrthe Spronck":MAILTO:m.s.c.spronck@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jeroen-keiren-an-expressi ve-timed-modal-mu-calculus-for-timed-automata/ END:VEVENT BEGIN:VEVENT UID:20240130T1008Z-1706609335.8363-EO-1496-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20240130T095639Z LAST-MODIFIED:20240130T095639Z DTSTART;TZID=Europe/Amsterdam:20240201T124500 DTEND;TZID=Europe/Amsterdam:20240201T133000 SUMMARY: Clemens Dubslaff: Explaining Control Strategies: Trees or Diagrams ? DESCRIPTION: Decision trees (DTs) are widely used to represent control stra tegies\, e.g.\, for machine learning classifiers or formal verification res ults. Decisions in DTs are based on expressive predicates\, but they seem t o not fully exploit their potential towards concise representations. One re ason is in their tree structure\, leading to isomorphic subtrees not being merged. Reduced ordered ...continue reading X-ALT-DESC;FMTTYPE=text/html:Decision trees (DTs) are widely used to re present control strategies\, e.g.\, for machine learning classifiers or for mal verification results. Decisions in DTs are based on expressive predicat es\, but they seem to not fully exploit their potential towards concise rep resentations. One reason is in their tree structure\, leading to isomorphic subtrees not being merged. Reduced ordered binary decision diagrams (BDDs) inherently support merging\, but allow only for reasoning on boolean varia bles with an ordering imposed. In this talk I will report on ongoing work c omparing decision trees and decision diagrams w.r.t. their capabilities to concisely represent control strategies and how to overcome their drawbacks.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Myrthe Spronck":MAILTO:m.s.c.spronck@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/clemens-dubslaff-explaini ng-control-strategies-trees-or-diagrams/ END:VEVENT BEGIN:VEVENT UID:20240123T1712Z-1706029977.6672-EO-1494-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20240123T165820Z LAST-MODIFIED:20240123T165820Z DTSTART;TZID=Europe/Amsterdam:20240125T124500 DTEND;TZID=Europe/Amsterdam:20240125T133000 SUMMARY: Maximilian Köhl: Most-Specific Verdictors for Imperfectly Observab le Systems DESCRIPTION: The observable behavior of a system usually carries crucial in formation about its internal state\, properties\, and potential future beha viors. In this talk\, I present a generic automata-theoretic synthesis appr oach to obtain most-specific verdicts from imperfect observations of an ong oing run of a system. Verdicts can be elements of any join-semilattice orde red by specificity. I show ...continue reading X-ALT-DESC;FMTTYPE=text/html:The observable behavior of a system usuall y carries crucial information about its internal state\, properties\, and p otential future behaviors. In this talk\, I present a generic automata-theo retic synthesis approach to obtain most-specific verdicts from imperfect ob servations of an ongoing run of a system. Verdicts can be elements of any j oin-semilattice ordered by specificity. I show that our approach covers exi sting work on runtime monitoring and fault diagnosis\, where verdicts indic ate the satisfaction or violation of properties\, or the occurrence of faul ts\, respectively. As an entirely novel application\, I present configurati on monitors\, yielding verdicts about possible configurations of a system. Furthermore\, I present empirical results obtained by configuration monitor ing on well-established configurable systems community benchmarks. If there is enough time\, I might also give a sneak peek on how the approach genera lizes to the real-time setting\, where the timing of events is often crucia l for monitoring and diagnosis\, but its precise assessment is hindered by timing imprecisions.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Myrthe Spronck":MAILTO:m.s.c.spronck@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/maximilian-kohl-most-spec ific-verdictors-for-imperfectly-observable-systems/ END:VEVENT BEGIN:VEVENT UID:20240116T1506Z-1705417578.5474-EO-1493-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20240116T150349Z LAST-MODIFIED:20240116T150349Z DTSTART;TZID=Europe/Amsterdam:20240118T124500 DTEND;TZID=Europe/Amsterdam:20240118T133000 SUMMARY: Bas Luttik: On the translation of pi-calculus into mCRL2 DESCRIPTION: In [1]\, Rob van Glabbeek proved that there does not exist a c ompositional translation of pi-calculus into CCS that is valid up to strong barbed bisimilarity\, but that there does exist a similarly valid composit ional transition if the communication facility is of CCS is upgraded to ACP -style communication. In my talk\, I will explore whether ...continue readi ng X-ALT-DESC;FMTTYPE=text/html:In [1]\, Rob van Glabbeek proved that ther e does not exist a compositional translation of pi-calculus into CCS that i s valid up to strong barbed bisimilarity\, but that there does exist a simi larly valid compositional transition if the communication facility is of CC S is upgraded to ACP-style communication. In my talk\, I will explore wheth er van Glabbeek’s translation can be modified to obtain a compositional tra nslation into mCRL2 that is valid up to strong barbed bisimilarity.
I should warn the audience in advance that this is very much work in progres s and that I will not present any results of my own that have been establis hed beyond reasonable doubt.
[1] Rob van Glabbeek. Comparing the Expr essiveness of the pi-calculus and CCS. ACM Transactions on Computational Lo gic\, Vol. 25\, No. 1\, Article 1\, 2023.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Myrthe Spronck":MAILTO:m.s.c.spronck@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/bas-luttik-on-the-transla tion-of-pi-calculus-into-mcrl2/ END:VEVENT BEGIN:VEVENT UID:20240108T1611Z-1704730310.7686-EO-1485-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20240108T160852Z LAST-MODIFIED:20240108T160852Z DTSTART;TZID=Europe/Amsterdam:20240111T124500 DTEND;TZID=Europe/Amsterdam:20240111T133000 SUMMARY: Michel Reniers: Supervisory Control Synthesis of Timed Automata Us ing Forcible Events DESCRIPTION: This paper presents an algorithm for synthesizing a supervisor for timed automata (TA) using the conventional supervisory control theory. The algorithm is directly applicable to TA without explicit transformation into finite automata\, and iteratively strengthens the guards of edges lab eled by controllable events and invariants of locations where the progressi on of time can be preempted ...continue reading X-ALT-DESC;FMTTYPE=text/html:This paper presents an algorithm for synth esizing a supervisor for timed automata (TA) using the conventional supervi sory control theory. The algorithm is directly applicable to TA without exp licit transformation into finite automata\, and iteratively strengthens the guards of edges labeled by controllable events and invariants of locations where the progression of time can be preempted by forcible events. The syn thesized supervisor\, also a TA\, is controllable\, maximally permissive\, and guarantees a non-blocking and safe supervised plant. The use of real-va lued clocks in TA makes it a practical modeling framework\, however\, the i nfinite state space brings challenges. The proposed algorithm addresses the se by providing a synthesis method that avoids the state-space explosion of finite automata and the loss of information that can result from abstracti on of real-time values.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Myrthe Spronck":MAILTO:m.s.c.spronck@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/michel-reniers-supervisor y-control-synthesis-of-timed-automata-using-forcible-events-2/ END:VEVENT BEGIN:VEVENT UID:20231207T1602Z-1701964946.2542-EO-1473-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20231207T154235Z LAST-MODIFIED:20231207T154235Z DTSTART;TZID=Europe/Amsterdam:20231214T124500 DTEND;TZID=Europe/Amsterdam:20231214T133000 SUMMARY: Jan Martens: Smaller proofs(?) for the language inequivalence of t wo DFAs DESCRIPTION: Deterministic finite automata (DFAs) are perhaps one of the si mplest models of computation. A classic result is that if two automata with n states are language inequivalent\, then there is a word of length at mos t n that is accepted by only one of the automata\, i.e. the word is disting uishing. In a sense this ...continue reading X-ALT-DESC;FMTTYPE=text/html:Deterministic finite automata (DFAs) are p erhaps one of the simplest models of computation. A classic result is that if two automata with n states are language inequivalent\, then there is a w ord of length at most n that is accepted by only one of the automata\, i.e. the word is distinguishing. In a sense this distinguishing word proofs/exp lains/witnesses why the DFAs are inequivalent. We are interested in explain ing DFA inequivalence with more sophisticated properties such as: "all word s with an even number of a's are accepted"\, which in turn is only satisfie d in one of the DFAs. In this talk we'll look at distinguishing DFAs which can act as such witnesses.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-martens-smaller-proof s-for-the-language-inequivalence-of-two-dfas/ END:VEVENT BEGIN:VEVENT UID:20231204T1915Z-1701717350.5893-EO-1461-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20231204T191546Z LAST-MODIFIED:20231207T154319Z DTSTART;TZID=Europe/Amsterdam:20231207T124500 DTEND;TZID=Europe/Amsterdam:20231207T133000 SUMMARY: Cancelled – Michel Reniers: Supervisory Control Synthesis of Timed Automata Using Forcible Events DESCRIPTION: This paper presents an algorithm for synthesizing a supervisor for timed automata (TA) using the conventional supervisory control theory. The algorithm is directly applicable to TA without explicit transformation into finite automata\, and iteratively strengthens the guards of edges lab eled by controllable events and invariants of locations where the progressi on of time can be preempted ...continue reading X-ALT-DESC;FMTTYPE=text/html:This paper presents an algorithm for synth esizing a supervisor for timed automata (TA) using the conventional supervi sory control theory. The algorithm is directly applicable to TA without exp licit transformation into finite automata\, and iteratively strengthens the guards of edges labeled by controllable events and invariants of locations where the progression of time can be preempted by forcible events. The syn thesized supervisor\, also a TA\, is controllable\, maximally permissive\, and guarantees a non-blocking and safe supervised plant. The use of real-va lued clocks in TA makes it a practical modeling framework\, however\, the i nfinite state space brings challenges. The proposed algorithm addresses the se by providing a synthesis method that avoids the state-space explosion of finite automata and the loss of information that can result from abstracti on of real-time values.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/michel-reniers-supervisor y-control-synthesis-of-timed-automata-using-forcible-events/ END:VEVENT BEGIN:VEVENT UID:20231127T0827Z-1701073636.3965-EO-1456-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20231127T082305Z LAST-MODIFIED:20231127T082305Z DTSTART;TZID=Europe/Amsterdam:20231201T140000 DTEND;TZID=Europe/Amsterdam:20231201T144500 SUMMARY: Erik Pardijs: Finite bases in the linear time-branching time spect rum with sequential composition and succesful termination. DESCRIPTION: The process algebra TSP extends the process algebra BCCSP with sequential composition and successful termination. For BCCSP it is known\, for all semantics in Van Glabbeek’s linear time-branching time spectrum of behavioural semantics\, whether a finite basis exists. In this thesis\, we study whether these semantics are finitely based over TSP. We discovered t hat none ...continue reading X-ALT-DESC;FMTTYPE=text/html:The process algebra TSP extends the proces s algebra BCCSP with sequential composition and successful termination. For BCCSP it is known\, for all semantics in Van Glabbeek’s linear time-branch ing time spectrum of behavioural semantics\, whether a finite basis exists. In this thesis\, we study whether these semantics are finitely based over TSP. We discovered that none of the congruences between Ready Simulation an d Completed Trace equivalence\, some of which were finitely based over BCCS P\, are finitely based\, regardless of alphabet cardinality. We also show t hat 2-Nested Simulation and Possible Futures equivalence still do not have a finite\, sound and ground-complete axiomatization. Furthermore\, we provi de a finite basis for Simulation equivalence with an infinite alphabet\, Tr ace equivalence with a non-singleton alphabet and Language equivalence for all alphabet cardinalities.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 3 GEO:51.447567;5.487443 ORGANIZER;CN="Bas Luttik":MAILTO:s.p.luttik@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/erik-pardijs-finite-bases -in-the-linear-time-branching-time-spectrum-with-sequential-composition-and -succesful-termination/ END:VEVENT BEGIN:VEVENT UID:20231128T1533Z-1701185581.8885-EO-1459-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20231128T152357Z LAST-MODIFIED:20231128T152529Z DTSTART;TZID=Europe/Amsterdam:20231201T140000 DTEND;TZID=Europe/Amsterdam:20231201T144500 SUMMARY: Jordi van Laarhoven: Formalising the State Machine Modelling Tool (SMMT) DESCRIPTION: Model-Driven (Software) Engineering (MDSE) is gaining populari ty in industry. More and more companies acknowledge the benefits of using M DSE to develop their software components. A company that exploits model-dri ven software engineering is Canon Production Printing (Venlo\, The Netherla nds). At Canon Production Printing\, the State Machine Modelling Tool (SMMT ) was developed to enable the modelling of ...continue reading X-ALT-DESC;FMTTYPE=text/html:Model-Driven (Software) Engineering (MDSE) is gaining popularity in industry. More and more companies acknowledge the benefits of using MDSE to develop their software components. A company tha t exploits model-driven software engineering is Canon Production Printing ( Venlo\, The Netherlands). At Canon Production Printing\, the State Machine Modelling Tool (SMMT) was developed to enable the modelling of software com ponents using state machines. In this graduation report\, we present SMMT a nd formally define a subset of the SMMT language. A translation from SMMT s pecifications to mCRL2 specifications is defined to allow for model checkin g on the SMMT specifications. We show that this translation can correctly g enerate mCRL2 specifications for the existing SMMT specifications at Canon Production Printing. Furthermore\, we show how the mCRL2 toolset can be use d to prove the correctness of SMMT specifications.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 13 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jordi-van-laarhoven-forma lising-the-state-machine-modelling-tool-smmt/ END:VEVENT BEGIN:VEVENT UID:20231127T1058Z-1701082696.8973-EO-1457-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20231127T105315Z LAST-MODIFIED:20231127T105315Z DTSTART;TZID=Europe/Amsterdam:20231130T124500 DTEND;TZID=Europe/Amsterdam:20231130T133000 SUMMARY: Thomas Neele: Characterising the Winning Strategies in Binary Pari ty Games with Propositional Logic DESCRIPTION: A parity game is an infinite-duration game between two players who pass a token around in a directed graph. The problem of “solving” a pa rity game (deciding which player wins a given node) is an interesting probl em since it is in UP and co-UP\, but not known to be in P. In a binary pari ty ...continue reading X-ALT-DESC;FMTTYPE=text/html:A parity game is an infinite-duration game between two players who pass a token around in a dir ected graph. The problem of "solving" a parity game (deciding which player wins a given node) is an interesting problem since it is in UP and co-UP\, but not known to be in P. In a binary parity game\, at least one player can make only binary choices. This allows us to characterise the set of winnin g strategies (i.e. along which of the two edges this player should move the token) in a propositional formula. In this talk\, I will discuss our progr ess trying to identify the sublogic that corresponds to such characterisati ons.
This is joint wor k with Tom Franken and Jan Friso Groote.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/thomas-neele-characterisi ng-the-winning-strategies-in-binary-parity-games-with-propositional-logic/ END:VEVENT BEGIN:VEVENT UID:20231122T1030Z-1700649009.7602-EO-1452-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20231122T101825Z LAST-MODIFIED:20231122T101825Z DTSTART;TZID=Europe/Amsterdam:20231123T124500 DTEND;TZID=Europe/Amsterdam:20231123T133000 SUMMARY: Pieter van Gelder: Probabilistic risk assessment of civil infrastr uctural systems DESCRIPTION: In this lecture I will discuss a number of methods and techniq ues for the probabilistic risk assessment of civil infrastructural systems\ , which includes extreme value statistics\, system decompositioning techniq ues\, and cost benefit analyses for risk-based optimisation. The flood defe nce system of the Netherlands will be used as a case study to illustrate th e applicability of ...continue reading X-ALT-DESC;FMTTYPE=text/html:In this lecture I will discuss a number of methods and techniques for the probabilistic risk assessment of civil infr astructural systems\, which includes extreme value statistics\, system deco mpositioning techniques\, and cost benefit analyses for risk-based optimisa tion. The flood defence system of the Netherlands will be used as a case st udy to illustrate the applicability of the above mentioned methods and tech niques.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/pieter-van-gelder-probabi listic-risk-assessment-of-civil-infrastructural-systems/ END:VEVENT BEGIN:VEVENT UID:20231207T1005Z-1701943502.6861-EO-1470-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20231207T100228Z LAST-MODIFIED:20231207T100301Z DTSTART;TZID=Europe/Amsterdam:20231123T100000 DTEND;TZID=Europe/Amsterdam:20231123T104500 SUMMARY: Jore Booy: Model-Based Mocking: Compositional Model-Based Testing for Microservices using the Axini Modeling Platform DESCRIPTION: Model-based testing is a compelling method for the integration testing of microservices. However\, when testing with a large number of se rvices\, state space explosions are a common problem. It is especially a pr oblem in model-based testing since the input-output conformance (ioco) rela tion is not compositional. We developed a novel and theoretically grounded testing method called ...continue reading X-ALT-DESC;FMTTYPE=text/html:Model-based testing is a compelling method for the integration testing of microservices. However\, when testing with a large number of services\, state space explosions are a common problem. I t is especially a problem in model-based testing since the input-output con formance (ioco) relation is not compositional. We developed a novel and the oretically grounded testing method called model-based mocking (MBM) to test microservice systems by mocking the underlying services of a SUT during te sting. MBM is sufficient to make component-wise testing compositional.
< p>We tested the MBM method using the Axini Modeling platform by inserting 2 0 mutants into an example microservice system. In our set of inserted bugs\ , MBM found more than half of the bugs faster compared to other methods and was slower for none of the bugs. CATEGORIES:MSc Defence LOCATION:Atlas 11.201 GEO:51.447551;5.487453 ORGANIZER;CN="Jeroen Keiren":MAILTO:j.j.a.keiren@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jore-booy-model-based-moc king-compositional-model-based-testing-for-microservices-using-the-axini-mo deling-platform/ END:VEVENT BEGIN:VEVENT UID:20231114T1145Z-1699962317.7722-EO-1445-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20231114T113613Z LAST-MODIFIED:20231114T113613Z DTSTART;TZID=Europe/Amsterdam:20231116T123000 DTEND;TZID=Europe/Amsterdam:20231116T133000 SUMMARY: Anna Stramaglia: Simplifying process parameters of unfolding algeb raic data types\, and Tom Franken: An Autonomous Data Language DESCRIPTION: Anna Stramaglia Title: Simplifying process parameters of unfol ding algebraic data types Abstract: In preparation for ICTAC 2023\, in this talk I will present the work done in collaboration with Jeroen Keiren and Thomas Neele. Complex abstract data types are often used to facilitate crea ting concise models of the behavior of realistic systems. However\, static analysis ...continue reading X-ALT-DESC;FMTTYPE=text/html:Anna Stramaglia
Title: Simplifying process parameters of unfolding algebraic data types
< em>Abstract:
In preparation for ICTAC 2023\, in this talk I wil l present the work done in collaboration with Jeroen Keiren and Thomas Neel e.
Complex abstract data types are often used to facilitate creating concise models of the behavior of realistic systems. However\, static analy sis techniques that aim to optimize such models often consider variables of complex types as a single indivisible unit. The use of complex data types thus negatively affects the optimizations that can be performed. To address this problem\, Groote and Lisser introduced a technique for flattening the structure of process parameters\, then implemented in mCRL2 in the tool lp sparunfold. We have extended the technique behind lpsparufold and implement ed the changes. In this talk I will first give some context\, I will then i ntroduce the original lpsparunfold technique by Groote and Lisser\, after I will describe our extensions and finally discuss the results of the applic ation of our extended technique on various specifications from different do mains.
Tom Franken
Title: An Autonomous Data Language
Abstract:
In the colloquium\, I will p ractice my talk to be given at ICTAC 2023. Therefore I will present AuDaLa\ , including a motivation\, some semantics and at least one example\, within 20-30 minutes. During ICTAC\, the abstract for the talk is the abstract of the paper:Nowadays\, the main advances in computational power are due to p arallelism. However\, most parallel languages have been designed with a foc us on processors and threads. This makes dealing with data and memory in pr ograms hard\, which distances the implementation from its original algorith m. We propose a new paradigm for parallel programming\, the data-autonomous paradigm\, where computation is performed by autonomous data elements. Pro grams in this paradigm are focused on making the data collaborate in a high ly parallel fashion. We furthermore present AuDaLa\, the first data autonom ous programming language\, and include an operational semantics. Programmin g in AuDaLa is very natural\, as illustrated by examples\, albeit in a styl e very different from sequential and contemporary parallel programming.
CATEGORIES:Colloquium ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/anna-stramaglia-simplifyi ng-process-parameters-of-unfolding-algebraic-data-types-and-tom-franken-an- autonomous-data-language/ END:VEVENT BEGIN:VEVENT UID:20231031T1106Z-1698750368.1607-EO-1435-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20231031T110059Z LAST-MODIFIED:20231031T110059Z DTSTART;TZID=Europe/Amsterdam:20231102T124500 DTEND;TZID=Europe/Amsterdam:20231102T133000 SUMMARY: Valentina Castiglioni: A formal framework for Cyber-Physical Syste ms DESCRIPTION: Cyber-Physical Systems (CPS) are characterised by the interact ion of various agents operating under highly changing and\, sometimes\, unp redictable environmental conditions. It is therefore fundamental to verify whether these systems are robust\, i.e.\, whether they are able to function correctly even in perturbed circumstances. In this talk I will recap the a chievements of my project on ...continue reading X-ALT-DESC;FMTTYPE=text/html:Cyber-Physical Systems (CPS) are character ised by the interaction of various agents operating under highly changing a nd\, sometimes\, unpredictable environmental conditions. It is therefore fu ndamental to verify whether these systems are robust\, i.e.\, whether they are able to function correctly even in perturbed circumstances.
In th is talk I will recap the achievements of my project on the development of a general formal framework and the tools necessary for the modelling\, analy sis and verification of CPS-like systems operating under uncertainty.
I will present the evolution sequence model for representing systems behav iour and the two temporal logics\, RobTL and DisTL\, introduced to specify robustness properties in various situations. I will also briefly introduce the "Software tool for the analsysis of robustness in the unkonw environmen t"\, that includes a domain specific language for the specification of syst ems and their properties\, and the model checkers for the two logics.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/valentina-castiglioni-a-f ormal-framework-for-cyber-physical-systems/ END:VEVENT BEGIN:VEVENT UID:20231023T0909Z-1698052176.0407-EO-1423-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20231023T090633Z LAST-MODIFIED:20231023T090633Z DTSTART;TZID=Europe/Amsterdam:20231026T150000 DTEND;TZID=Europe/Amsterdam:20231026T154500 SUMMARY: Isabelle Cooijmans: Verification of EULYNX light signal using mCRL 2 and comparing with auto-translated model DESCRIPTION: This thesis presents the light signal case study as part of th e FormaSig project\, which aims to use formal methods to support the develo pment of EULYNX – a European initiative to standardise interfaces of signal ling systems. EULYNX specifies its interfaces using SysML\, which is a semi -formal modelling language. Previous research has developed a formalisation for ...continue reading X-ALT-DESC;FMTTYPE=text/html:Labeled continuous-time Markov chains (CTM Cs) describe processes subject to random timing and partial observability. In applications such as runtime monitoring or calculation of remaining usef ul life\, we must incorporate past observations. The timing of these observ ations matters but they may be uncertain.
We consider a setting in wh ich we are given a sequence of imprecisely timed labels called the evidence . The problem is to compute reachability probabilities\, which we condition on this evidence.
In this talk\, I will present a method that solves this problem by (1) unfolding the CTMC into a continuous Markov decision p rocess (MDP)\, (2) abstracting the continuous MDP into a finite interval MD P and (3) applying an iterative refinement scheme to obtain bounds on the c onditional probabilities in the CTMC.
(joint work with Thom Badings\, Sebastian Junges\, Nils Jansen and Marielle Stoelinga)
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/matthias-volk-continuous- time-markov-chains-with-imprecisely-timed-observations/ END:VEVENT BEGIN:VEVENT UID:20231023T0909Z-1698052176.0483-EO-1421-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20231023T090009Z LAST-MODIFIED:20231023T090250Z DTSTART;TZID=Europe/Amsterdam:20231023T160000 DTEND;TZID=Europe/Amsterdam:20231023T173000 SUMMARY: Mark Bouwman: Supporting Railway Standardisation with Formal Verif ication DESCRIPTION: On October 23\, 2023\, Mark Bouwman will defend his thesis tit led ‘Supporting Railway Standardisation with Formal Verification’\, which i s available via https://research.tue.nl/en/publications/supporting-railway- standardisation-with-formal-verification. The defence marks a milestone in the FormaSig project\, a collaboration between TU/e\, UT\, ProRail and Deut sche Bahn. X-ALT-DESC;FMTTYPE=text/html:On October 23\, 2023\, Mark Bouwman will d efend his thesis titled ‘Supporting Railway Standardisation with Formal Ver ification’\, which is available via https://research.tue.nl/en/publications /supporting-railway-standardisation-with-formal-verification. The defence m arks a milestone in the FormaSig project\, a collaboration between TU/e\, U T\, ProRail and Deutsche Bahn.
CATEGORIES:PhD Defence LOCATION:Atlas 0.710 GEO:51.446751;5.486286 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/1421/ END:VEVENT BEGIN:VEVENT UID:20231017T1107Z-1697540834.0862-EO-1412-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20231017T105727Z LAST-MODIFIED:20231017T105727Z DTSTART;TZID=Europe/Amsterdam:20231019T124500 DTEND;TZID=Europe/Amsterdam:20231019T133000 SUMMARY: Flip van Spaendonck: (Semi-)Automate Extraction of Behavioral Mode ls from C++ Code DESCRIPTION: Accurate behavioral models of software systems can be incredib ly useful thanks to the vast array of model based techniques that exist\, e .g. automated testing through model based testing\, verification of require ments using modal formulas\, or providing visual insight using state space visualization techniques. However\, acquiring such models can be quite diff icult. We propose a set ...continue reading X-ALT-DESC;FMTTYPE=text/html:Accurate behavioral models of software sys tems can be incredibly useful thanks to the vast array of model based techn iques that exist\, e.g. automated testing through model based testing\, ver ification of requirements using modal formulas\, or providing visual insigh t using state space visualization techniques. However\, acquiring such mode ls can be quite difficult. We propose a set of transformation rules for the C++ language that allows for the generation of an equivalent behavioral mo del and showcase our current progress on a tool that allows for the automat ic transformation of c++ code into behavioral models using these transforma tion rules. We note that\, whilst the given transformation rules are specif ic to C++\, they should be able to be easily transferred to other object or iented languages as well.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/flip-van-spaendonck-semi- automate-extraction-of-behavioral-models-from-c-code/ END:VEVENT BEGIN:VEVENT UID:20231009T0854Z-1696841683.5633-EO-1411-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20231009T084009Z LAST-MODIFIED:20231009T084009Z DTSTART;TZID=Europe/Amsterdam:20231012T124500 DTEND;TZID=Europe/Amsterdam:20231012T133000 SUMMARY: Jan Friso Groote: Numerically solving Real Equation Systems DESCRIPTION: The quantitative modal mu-calculus is equal to the modal mu-ca lculus except that formulas yield real values including (-)infinity instead of true or false. Quantitative modal formulas can be translated to paramet erised real equation systems (PRESs)\, and subsequently to real equation sy stems (RESs)\, which is quite similar as translating to PBESs and BESs in c ase of ...continue reading X-ALT-DESC;FMTTYPE=text/html:The quantitative modal mu-calculus is equa l to the modal mu-calculus except that formulas yield real values including (-)infinity instead of true or false. Quantitative modal formulas can be t ranslated to parameterised real equation systems (PRESs)\, and subsequently to real equation systems (RESs)\, which is quite similar as translating to PBESs and BESs in case of ordinary formula.
Gauss elimination is a c omplete method to solve RESs\, but it is so inefficient that only small RES s can be solved. There is a numerical algorithm\, which will be explained\, which appears very efficient. However\, the algorithm cannot deal with sit uations where solutions go to infinity.
During this presentation I wi ll elaborate on this and indicate where I got stuck.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-friso-groote-numerica lly-solving-real-equation-systems/ END:VEVENT BEGIN:VEVENT UID:20231004T0851Z-1696409510.9286-EO-1410-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20231004T084736Z LAST-MODIFIED:20231004T084736Z DTSTART;TZID=Europe/Amsterdam:20231005T124500 DTEND;TZID=Europe/Amsterdam:20231005T133000 SUMMARY: Kevin Jilissen: Behavioral comparison of SysML and Dezyne models DESCRIPTION: In this talk\, I will focus on a work-in-progress behavioral a bstraction enabling the comparison of two modelling methods. The first meth od is an adaption of the current modelling methodology at Rijkswaterstaat f or tunnel control systems based on SysML Activity Diagrams\, as introduced in my previous talk. The second method relies on the Dezyne specification l anguage ...continue reading X-ALT-DESC;FMTTYPE=text/html:In this talk\, I wil l focus on a work-in-progress behavioral abstraction enabling the compariso n of two modelling methods. The first method is an adaption of the current modelling methodology at Rijkswaterstaat for tunnel control systems based o n SysML Activity Diagrams\, as introduced in my previous talk. The second m ethod relies on the Dezyne specification language by Verum. Ideally\, both modelling methods should describe the same system when starting from some s hared specification. As these modelling methods have totally different char acteristics for a.o. atomicity in the resulting state space\, the presented abstraction enables reasoning about the equivalence of the behavior contai ned in both models.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/kevin-jilissen-behavioral -comparison-of-sysml-and-dezyne-models/ END:VEVENT BEGIN:VEVENT UID:20230927T0713Z-1695798819.9192-EO-1409-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230927T071251Z LAST-MODIFIED:20230927T071251Z DTSTART;TZID=Europe/Amsterdam:20230928T093000 DTEND;TZID=Europe/Amsterdam:20230928T103000 SUMMARY: Edward Liem: Extraction of Invariants in Parameterised Boolean Equ ation Systems DESCRIPTION: Parameterised Boolean Equation Systems (PBESs) are used to exp ress and solve various model checking and equivalence checking problems. Ho wever\, it may not always be efficient\, or even possible\, to find a solut ion to PBESs since they may encode undecidable problems. One particular tec hnique towards finding a solution to a PBES is the concept of exploiting .. .continue reading X-ALT-DESC;FMTTYPE=text/html:Parameterised Boolean Equation Systems (PB ESs) are used to express and solve various model checking and equivalence c hecking problems. However\, it may not always be efficient\, or even possib le\, to find a solution to PBESs since they may encode undecidable problems . One particular technique towards finding a solution to a PBES is the conc ept of exploiting global PBES invariants. Although invariants have been stu died extensively\, there is a lack of research towards invariant discovery and exploitation in PBESs. Our paper presents PBES invariant extraction tec hniques inspired from various concepts found in program verification litera ture well as provide new conditions for invariance properties. We also pres ent a novel graph structure\, namely relevancy graphs\, which characterize relevant predicate variable instances of instantiated PBES equations. Using relevancy graphs\, we illustrate how invariants interact with PBESs as wel l as provide an alternative criteria to proving the PBES global invariant c ondition in simple functions.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 14 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/edward-liem-extraction-of -invariants-in-parameterised-boolean-equation-systems/ END:VEVENT BEGIN:VEVENT UID:20230912T1021Z-1694514113.4467-EO-1407-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230912T100529Z LAST-MODIFIED:20230912T100529Z DTSTART;TZID=Europe/Amsterdam:20230914T124500 DTEND;TZID=Europe/Amsterdam:20230914T133000 SUMMARY: Tim Willemse: Model-Driven Engineering meets Model-Based Testing DESCRIPTION: In this talk\, I will focus on a connection between stable-fai lures refinement and the ioco conformance relation. Both behavioural relati ons underlie methodologies that have gained traction in industry: stable-fa ilures refinement is used in several commercial Model-Driven Engineering to ol suites\, whereas the ioco conformance relation is used in Model-Based Te sting tools. Refinement-based Model-Driven Engineering approaches promise . ..continue reading X-ALT-DESC;FMTTYPE=text/html:In this talk\, I will focus on a connectio n between stable-failures refinement and the ioco conformance relation. Bot h behavioural relations underlie methodologies that have gained traction in industry: stable-failures refinement is used in several commercial Model-D riven Engineering tool suites\, whereas the ioco conformance relation is us ed in Model-Based Testing tools.
Refinement-based Model-Driven Engine ering approaches promise to generate executable code from high-level models \, thus guaranteeing that the code upholds specified behavioural contracts. Manual testing\, however\, is still required to gain confidence that the m odel-to-code transformation and the execution platform do not lead to unexp ected contract violations.
We identify conditions under which also th is last step in the design methodology can be automated using the ioco conf ormance relation and the associated tools.
Joint work with Flip van S paendonck
CATEGORIES:Colloquium LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tim-willemse-model-driven -engineering-meets-model-based-testing/ END:VEVENT BEGIN:VEVENT UID:20230904T1500Z-1693839640.7555-EO-1396-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230904T144944Z LAST-MODIFIED:20230904T144944Z DTSTART;TZID=Europe/Amsterdam:20230907T140000 DTEND;TZID=Europe/Amsterdam:20230907T150000 SUMMARY: Myrthe Spronck: Fairness Assumptions in the Modal mu-Calculus DESCRIPTION: The modal μ-calculus is a highly expressive logic\, but its fo rmulae are often hard to understand. We have tools for testing if a model s atisfies a model μ-calculus formula\, but if we are unsure of what the form ula expresses we cannot draw definite conclusions from the results. To miti gate the difficulties in designing μ-calculus formulae\, ...continue readin g X-ALT-DESC;FMTTYPE=text/html:The modal μ-calculus is a highly expressiv e logic\, but its formulae are often hard to understand. We have tools for testing if a model satisfies a model μ-calculus formula\, but if we are uns ure of what the formula expresses we cannot draw definite conclusions from the results. To mitigate the difficulties in designing μ-calculus formulae\ , property specification patterns have been designed to help researchers ex press common properties in the μ-calculus. However\, existing translations of these patterns to the modal μ-calculus only allow for the inclusion fair ness assumptions to a very limited degree\, even though fairness assumption s are very useful when model checking. Fairness assumptions allow the resea rcher to specify that certain types of property violations that may exist i n the model are unrealistic and should not be considered when determining i f the property is satisfied. This is often required because most of the tim e\, when researchers model systems they abstract away from certain details such as the way scheduling is done\, which then results in scenarios being represented in the model that would never occur in reality. There exists\, therefore\, a need for a standard and proven way of including a variety of fairness assumptions in modal μ-calculus formulae for properties. We extend the existing translations from the property specification patterns to the modal μ-calculus with ways to incorporate three common fairness assumptions : weak fairness\, strong fairness and fair reachability (also known as ∞-fa irness or hyperfairness) of the actions in a model. We also discuss other f airness assumptions to a lesser extend\, including unconditional fairness o f actions and weak fairness\, strong fairness and fair reachability of part s of a model other than actions. When it comes to the patterns\, we start w ith a detailed discussion on the global response pattern\, which is one of the most commonly occurring ones. We then generalise our approach to cover other patterns as well. Correctness proofs are included for both the global response formulae and the generalised formulae. We conclude we a brief dis cussion of how the formulae we have presented in this thesis can be used in the model checking toolset mCRL2.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 14 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/myrthe-spronck-fairness-a ssumptions-in-the-modal-mu-calculus/ END:VEVENT BEGIN:VEVENT UID:20230906T0918Z-1693991908.6653-EO-1399-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230906T090528Z LAST-MODIFIED:20230906T090528Z DTSTART;TZID=Europe/Amsterdam:20230907T124500 DTEND;TZID=Europe/Amsterdam:20230907T133000 SUMMARY: Anton Wijs: The Hitchhiking Algorithm for Massively Parallel LTL M odel Checking DESCRIPTION: Efficient algorithms have been developed to model check LTL fo rmulae on-the-fly\, such as the well-known Nested Depth-First Search\, whic h uses a depth-first search (DFS) strategy. However\, in some settings\, su ch as when considering distributed model checking on a cluster\, or many-co re model checking using a Graphics Processing Unit (GPU)\, Breadth-First Se arch (BFS) is a more ...continue reading X-ALT-DESC;FMTTYPE=text/html:Efficient algorithms have been developed t o model check LTL formulae on-the-fly\, such as the well-known Nested Depth -First Search\, which uses a depth-first search (DFS) strategy. However\, i n some settings\, such as when considering distributed model checking on a cluster\, or many-core model checking using a Graphics Processing Unit (GPU )\, Breadth-First Search (BFS) is a more natural choice\, at least for basi c reachability analysis. LTL verification\, however\, requires the detectio n of `accepting' cycles\, and BFS is not very suitable to detect these on-t he-fly. I will propose a new algorithm\, the Hitchhiking algorithm\, for mo del checking LTL formulae with a BFS-like strategy. It builds on the existi ng Piggyback algorithm from 2012\, but whereas that algorithm is incomplete \, the Hitchhiking algorithm is complete.
CATEGORIES:Colloquium LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/anton-wijs-the-hitchhikin g-algorithm-for-massively-parallel-ltl-model-checking/ END:VEVENT BEGIN:VEVENT UID:20230814T1129Z-1692012560.8296-EO-1385-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230814T112547Z LAST-MODIFIED:20230814T112547Z DTSTART;TZID=Europe/Amsterdam:20230828T100000 DTEND;TZID=Europe/Amsterdam:20230828T103000 SUMMARY: Gijs Leemrijse: Towards relaxed memory semantics for the Autonomou s Data Language DESCRIPTION: On 28 August at 10:00 in MF14\, Gijs Leemrijse will defend his MSc thesis titled “Towards relaxed memory semantics for the Autonomous Dat a Language”. This work presents an alternative operational semantics for th e Autonomous Data Language (AuDaLa) with relaxed memory consistency and inc oherent memory. We show how the memory operations of our semantics can be . ..continue reading X-ALT-DESC;FMTTYPE=text/html:On 28 August at 10:00 in MF14\, Gijs Leemr ijse will defend his MSc thesis titled "Towards relaxed memory semantics fo r the Autonomous Data Language".
This work presents an alternative op erational semantics for the Autonomous Data Language (AuDaLa) with relaxed memory consistency and incoherent memory. We show how the memory operations of our semantics can be safely mapped onto the NVIDIA PTX virtual ISA and demonstrate that our semantics performs faster than the original when execu ting AuDaLa programs on GPUs. We translate our operational semantics into a n axiomatic memory consistency model and formally check\, for a bounded pro gram size\, its correspondence with PTX’s memory consistency model using th e Alloy model finder. We conclude by presenting AuDaLaC\, our compiler targ eting the CUDA platform\, with which we explore several different strategie s to compile AuDaLa programs. We demonstrate in several case studies that A uDaLa implementations can perform faster than sequential implementations. p> CATEGORIES:MSc Defence LOCATION:MetaForum MF 14 GEO:51.447532;5.487437 ORGANIZER;CN="Thomas Neele":MAILTO:t.s.neele@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/gijs-leemrijse-towards-re laxed-memory-semantics-for-the-autonomous-data-language/ END:VEVENT BEGIN:VEVENT UID:20230710T1229Z-1688992140.1357-EO-1379-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230710T120942Z LAST-MODIFIED:20230710T120942Z DTSTART;TZID=Europe/Amsterdam:20230713T124500 DTEND;TZID=Europe/Amsterdam:20230713T133000 SUMMARY: Rick Erkens: Automaton-based Techniques for Optimized Term Rewriti ng DESCRIPTION: After five years of PhD research I am finishing my thesis on a lgorithms for fast term rewriting. In this talk I will look back on the pro ject\, summarize the results\, and list the open problems X-ALT-DESC;FMTTYPE=text/html:
After five years of PhD research I am fini shing my thesis on algorithms for fast term rewriting. In this talk I will look back on the project\, summarize the results\, and list the open proble ms
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rick-erkens-automaton-bas ed-techniques-for-optimized-term-rewriting/ END:VEVENT BEGIN:VEVENT UID:20230707T0805Z-1688717128.6517-EO-1377-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230707T080450Z LAST-MODIFIED:20230707T080540Z DTSTART;TZID=Europe/Amsterdam:20230707T130000 DTEND;TZID=Europe/Amsterdam:20230707T133000 SUMMARY: Tim Beurskens: Formal Verification of Safety Properties in Automot ive Systems DESCRIPTION: As the automotive industry transitions towards model-based dev elopment\, the need for model-based verification arises. Although recognize d institutes such as ISO and IEC recognize the benefits of formal analysis in the development of safety-critical E/E systems\, specific standards or g uidelines on these practices are mostly absent. There are several toolsets available for formal verification purposes. These ...continue reading X-ALT-DESC;FMTTYPE=text/html:As the automotive industry transitions towards model-based development\, the need for model-based verification ari ses. Although recognized institutes such as ISO and IEC recognize the benef its of formal analysis in the development of safety-critical E/E systems\, specific standards or guidelines on these practices are mostly absent. Ther e are several toolsets available for formal verification purposes. These to olsets often use a niche modeling syntax which could hinder industry adapta tion. By bridging the gap between widely used modeling software such as Sim ulink and verification tools we could enable the adoption of formal methods in the automotive or other safety-critical industries.
This work explored how formal verification can be applied to hierarchical Simul ink and Simulink Stateflow models. By dividing the models into separate sub systems we can construct large hierarchical models in mCRL2 syntax\, simila r to the original hierarchy in Simulink. Several blocks in the Simulink blo ck library have been translated to a functionally similar variant in mCRL2. Instances of these implementations can be placed in dataflow diagrams to o btain complex behaviour from a collection of simple processes. This process has been partially automated using a code generator which interprets mCRL2 design files as templates. This can reduce code repetition and makes the d esign process easier.
The approach has been demonstrated on a small toy example of the Collatz sequence\, as well as a model from Mathw orks’ examples repository\, modeling a traffic intersection with two traffi c lights. It could be demonstrated that the two variants of the Collatz mod el both converged to 1 for any of the given initial conditions.
The t raffic light example model featured both dataflow blocks\, virtual (subsyst em) blocks and hierarchical Stateflow charts in which message queues are us ed to communicate between charts. Most effort in translating this model wen t into the Stateflow controller model\, as this model used several Stateflo w features (message queues and hierarchical charts) with limited documentat ion. As there is no formal definition of a dataflow or Stateflow diagram\, the translation to mCRL2 has no formal basis. It can therefore not be demon strated that the translation results in functionally equivalent models comp ared to the Simulink sources.
A SMT-based scheduling tool has been used to overcome the issue of incorrect schedules in dataflow diagrams due to an incorrect port ordering. This tool computes a satisfying schedule ba sed on a simplified representation of the dataflow diagram. The resulting s chedule can be imported in the template preprocessor to re-order the output ports.
In this talk\, I will give the intuition a nd outline of how to prove our new data-autonomous programming language AuD aLa Turing Complete. While doing that\, I shall also discuss relevant aspec ts of the semantics\, including the basics\, null-elements and commands.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tom-franken-audala-turing -completeness-semantics/ END:VEVENT BEGIN:VEVENT UID:20230620T1107Z-1687259225.0075-EO-1374-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230620T110509Z LAST-MODIFIED:20230620T110509Z DTSTART;TZID=Europe/Amsterdam:20230622T124500 DTEND;TZID=Europe/Amsterdam:20230622T133000 SUMMARY: Bas Luttik: Verifying mutual exclusion algorithms: dropping the at omicity assumption DESCRIPTION: When formally verifying the correctness of mutual exclusion al gorithms it is often assumed that interaction with the shared registers (i. e.\, reads and writes) are atomic. For instance\, it is well-known that the correctness of Peterson’s algorithm relies on the atomicity assumption. Al ready in 1986\, however\, Lamport argued that implementing atomic interacti on with shared registers basically ...continue reading X-ALT-DESC;FMTTYPE=text/html:When formally verifying the correctness of mutual exclusion algorithms it is often assumed that interaction with the shared registers (i.e.\, reads and writes) are atomic. For instance\, it is well-known that the correctness of Peterson’s algorithm relies on the atom icity assumption. Already in 1986\, however\, Lamport argued that implement ing atomic interaction with shared registers basically requires mutual excl usion at the lower level. To get out of this chicken-and-egg problem\, we n eed mutual exclusion algorithms that do not rely on the atomicity assumptio n.
Myrthe Spronck and I have developed a method to verify mutual excl usion algorithms without the atomicity assumption with the mCRL2 toolset. W e used our method to verify several mutual exclusion algorithms for which i t was claimed (in print) that they were robust for non-atomic register inte ractions. We found that several of these algorithms are actually not as rob ust as they were claimed to be.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/bas-luttik-verifying-mutu al-exclusion-algorithms-dropping-the-atomicity-assumption/ END:VEVENT BEGIN:VEVENT UID:20230612T0816Z-1686557776.9103-EO-1373-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230612T081610Z LAST-MODIFIED:20230612T081610Z DTSTART;TZID=Europe/Amsterdam:20230615T124500 DTEND;TZID=Europe/Amsterdam:20230615T133000 SUMMARY: Jos Baeten: Parallel Pushdown Automata and Commutative Context-Fre e Grammars in Bisimulation Semantics DESCRIPTION: A classical theorem states that the set of languages given by a pushdown automaton coincides with the set of languages given by a context -free grammar. In a recent article\, Bas Luttik and I proved the pendant of this theorem in a setting with interaction: the set of processes given by a pushdown automaton coincides with ...continue reading X-ALT-DESC;FMTTYPE=text/html:A classical theorem states that the set of languages given by a pushdown automaton coincides with the set of language s given by a context-free grammar. In a recent article\, Bas Luttik and I p roved the pendant of this theorem in a setting with interaction: the set of processes given by a pushdown automaton coincides with the set of processe s given by a finite guarded recursive specification over a process algebra with actions\, choice\, and sequencing with sequential value passing. In a current paper\, we look what happens if we consider parallel pushdown autom ata instead of pushdown automata\, and a process algebra with parallelism i nstead of sequencing.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jos-baeten-parallel-pushd own-automata-and-commutative-context-free-grammars-in-bisimulation-semantic s/ END:VEVENT BEGIN:VEVENT UID:20230606T1146Z-1686052000.0735-EO-1372-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230606T113746Z LAST-MODIFIED:20230606T113746Z DTSTART;TZID=Europe/Amsterdam:20230608T124500 DTEND;TZID=Europe/Amsterdam:20230608T133000 SUMMARY: Jan Martens: Using the Strong Exponential Time Hypothesis (SETH) t o show the hardness of PTIME problems. DESCRIPTION: Unconditional lowerbounds on run-time complexity are challengi ng. For instance\, the best known lowerbound for CNF-SAT remains only linea r\, despite considerable research effort. Nevertheless\, we consider NP-har dness a reliable indication that we should not search for a polynomial time algorithms. Having a similar framework indicating the intrinsic difficulty of problems within PTIME\, such as linear\, quadratic\, ...continue readin g X-ALT-DESC;FMTTYPE=text/html:Unconditional lowerbounds on run-time comp lexity are challenging. For instance\, the best known lowerbound for CNF-SA T remains only linear\, despite considerable research effort. Nevertheless\ , we consider NP-hardness a reliable indication that we should not search f or a polynomial time algorithms. Having a similar framework indicating the intrinsic difficulty of problems within PTIME\, such as linear\, quadratic\ , cubic\, etc.\, would be highly valuable.
In this talk\, we will dem onstrate how the SETH (Strong Exponential Time Hypothesis)\, which roughly states that CNF-SAT cannot be solved more efficiently than brute force\, pr ovides such hardness-results. Specifically\, we demonstrate how SETH implie s that deciding the emptiness of the intersection of two deterministic auto mata is inherently quadratic. This result also directly suggests that solvi ng simulation equivalence for LTSs in subquadratic time is unlikely.
[1] Virginia Vassilevska Williams. "Hardness of easy problems: Basing hardn ess on popular conjectures such as the strong exponential time hypothesis ( invited talk)." - IPEC 2015
[2] Michael Wehar - "On the complexity o f intersection non-emptiness problems" PhD. Thesis 2016\, University at Buf falo
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-martens-using-the-str ong-exponential-time-hypothesis-seth-to-show-the-hardness-of-ptime-problems / END:VEVENT BEGIN:VEVENT UID:20230530T0846Z-1685436404.698-EO-1367-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230530T083957Z LAST-MODIFIED:20230530T083957Z DTSTART;TZID=Europe/Amsterdam:20230601T124500 DTEND;TZID=Europe/Amsterdam:20230601T133000 SUMMARY: Herman Geuvers: Directed Hennessy-Milner theorems DESCRIPTION: This is joint work with Anton Golov (RU) Labelled transitions systems can be studied both in terms of modal logic and in terms of bisimul ation. These two notions are connected by so-called Hennessy-Milner theorem s\, that show that states are bisimilar precisely when they satisfy the sam e formulas in some modal logic\, in other words\, when ...continue reading X-ALT-DESC;FMTTYPE=text/html:This is joint work with Anton Golov (RU) L abelled transitions systems can be studied both in terms of modal logic and in terms of bisimulation. These two notions are connected by so-called Hen nessy-Milner theorems\, that show that states are bisimilar precisely when they satisfy the same formulas in some modal logic\, in other words\, when they have the same theory according to this modal logic.
We introduce a directed version of such theorems\, from which the original results foll ow. To this end\, we introduce positive modal logics and directed notions o f bisimulation\, and then show that the theory of p in this positive logic is included in the theory of q precisely when p is directed bisimilar to q.
We develop these theories in two settings\, that of Hennessy-Milner Logic and strong bisimulation\, and that of Hennessy-Milner Logic with Unti l
(HMLU) and branching bisimulation. In the process\, we show that ev ery HMLU formula is equivalent to one satisfying a certain positivity prope rty. This gives rise to a sublogic of HMLU that is equally expressive but e asier to reason about.
By formulating our constructions in terms of a partness (the complement of bisimulation)\, we moreover find that all our d irected Hennessy-Milner theorems can be proven by structural induction. Thi s results in simpler proofs than one would obtain with traditional approach es.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/herman-geuvers-directed-h ennessy-milner-theorems/ END:VEVENT BEGIN:VEVENT UID:20230522T1610Z-1684771859.3452-EO-1366-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230522T160436Z LAST-MODIFIED:20230522T160436Z DTSTART;TZID=Europe/Amsterdam:20230525T124500 DTEND;TZID=Europe/Amsterdam:20230525T133000 SUMMARY: Anna Stramaglia: Simplifying process parameters of unfolding algeb raic data types DESCRIPTION: Complex abstract data types are often used to facilitate creat ing concise models of the behavior of realistic systems. However\, static a nalysis techniques that aim to optimize such models often consider variable s of complex types as a single indivisible unit. The use of complex data ty pes thus negatively affects the optimizations that can be performed. To ... continue reading X-ALT-DESC;FMTTYPE=text/html:Complex abstract data types are often used
to facilitate creating concise models of the behavior of realistic systems
. However\, static analysis techniques that aim to optimize such models oft
en consider variables of complex types as a single indivisible unit. The us
e of complex data types thus negatively affects the optimizations that can
be performed.
To address this problem\, Groote and Lisser introduced a
technique for flattening the structure of process parameters\, then implem
ented in mCRL2 in the tool lpsparunfold. We have extended the technique beh
ind lpsparufold and implemented the changes.
In this talk I will first
introduce the original lpsparunfold technique by Groote and Lisser (with a
n example)\, then describe our extensions (with examples) and finally discu
ss the results of the application of our extended technique on various spec
ifications from different domains.
This is joint work with Jeroen Kei ren and Thomas Neele.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/anna-stramaglia-simplifyi ng-process-parameters-of-unfolding-algebraic-data-types/ END:VEVENT BEGIN:VEVENT UID:20230515T0727Z-1684135631.6471-EO-1358-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230515T071345Z LAST-MODIFIED:20230515T071345Z DTSTART;TZID=Europe/Amsterdam:20230525T103000 DTEND;TZID=Europe/Amsterdam:20230525T113000 SUMMARY: Sky Sarah van Grieken: Replicating Experiments and Improving Algor ithms for Exact DFA Identification DESCRIPTION: In process mining\, grammatical inference is used to extract i nformation from large amounts of data. One problem in grammatical inference is exact DFA identification: finding a smallest DFA that accepts and rejec ts given words. Empirical experiments are used to compare new algorithms wi th the state of the art for exact DFA identification. To investigate to ... continue reading X-ALT-DESC;FMTTYPE=text/html:In process mining\, grammatical inference is used to extract information from large amounts of data. One problem in g rammatical inference is exact DFA identification: finding a smallest DFA th at accepts and rejects given words. Empirical experiments are used to compa re new algorithms with the state of the art for exact DFA identification. T o investigate to what extend the research field of formal verification is a ffected by the replication crisis\, we replicate the existing experiments t o see if their results are valid. We also run new experiments for exact DFA identification to test if the assumptions in existing papers are valid and to test if the performance of SAT and SMT solvers with new combinations of existing algorithms and symmetry breaking predicates can be improved. This thesis provides a program with all the state-of-the-art algorithms and sym metry breaking predicates of exact DFA identification\, with which our expe riments are easy to reproduce. With these experiments\, we show that the re plication crisis also affects the field of formal verification and replicat ing experiments can be made more attractive by also investigating new resea rch questions with the replicated code.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Thomas Neele":MAILTO:t.s.neele@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/sky-sarah-van-grieken-rep licating-experiments-and-improving-algorithms-for-exact-dfa-identification/ END:VEVENT BEGIN:VEVENT UID:20230515T1403Z-1684159433.6927-EO-1360-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230515T135754Z LAST-MODIFIED:20230515T135754Z DTSTART;TZID=Europe/Amsterdam:20230515T150000 DTEND;TZID=Europe/Amsterdam:20230515T154500 SUMMARY: Maarten Visscher: Formal verification on the Maeslant Barrier Loco mobile software DESCRIPTION: The Maeslant Barrier Locomobile software controls the barrier arms of the Maeslant storm surge barrier. The actual software controller of the locomobile has been literally modelled in mCRL2. The software was desc ribed in a document of over 500 pages. Subsequently\, 17 properties have be en extracted from the documentation of Rijkswaterstaat\, modelled as modal formulas verified ...continue reading X-ALT-DESC;FMTTYPE=text/html:The Maeslant Barrier Locomobile software c ontrols the barrier arms of the Maeslant storm surge barrier. The actual so ftware controller of the locomobile has been literally modelled in mCRL2. T he software was described in a document of over 500 pages. Subsequently\, 1 7 properties have been extracted from the documentation of Rijkswaterstaat\ , modelled as modal formulas verified on the model. The verification includ es the full behaviour\, including well defined erroneous behaviour\, for in stance due to faulty sensors\, except for a limited number of situations th at caused the statespace to grow too much.
This project clearly shows that it is possible to verify actual control software that occur storm bar riers.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 14 GEO:51.447532;5.487437 ORGANIZER;CN="Jan Friso Groote":MAILTO:j.f.groote@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/maarten-visscher-formal-v erification-on-the-maeslant-barrier-locomobile-software/ END:VEVENT BEGIN:VEVENT UID:20230508T0834Z-1683534875.6826-EO-1354-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230508T082603Z LAST-MODIFIED:20230508T082603Z DTSTART;TZID=Europe/Amsterdam:20230511T124500 DTEND;TZID=Europe/Amsterdam:20230511T133000 SUMMARY: Jeroen Keiren: Model checking in the context of digital twins DESCRIPTION: A In the setting of digital twins\, real world systems and vir tual models are kept in sync. If the implementation and the models share a common source of truth\, such as a low-code model\, this allows for a very tight integration of the different aspects of digital twins. In a collabora tion with\, among others\, Eindhoven ...continue reading X-ALT-DESC;FMTTYPE=text/html:A In the setting of digital twins\, real w orld systems and virtual models are kept in sync. If the implementation and the models share a common source of truth\, such as a low-code model\, thi s allows for a very tight integration of the different aspects of digital t wins. In a collaboration with\, among others\, Eindhoven University\, Cordi s SUITE and TNO ESI\, an environment was developed in which Cordis SUITE’s low code models are used to as ground truth. On the one hand\, these models are used to generate code for machine control applications\, with the nece ssary instrumentation to observe the state of the system and the data for d igital twinning. On the other hand\, the models are used for formal verific ation purposes.
In this talk I will describe the general overview of this digital twin environment. Subsequently\, I will focus on the applicati on of the mCRL2 model checker for the formal verification of the low code m odels\, and how the verification results can be fed back into the digital t win environment.
The results presented in this talk are part of the I TEA3 MACHINAIDE and the OPZuid Verification Base Remote & Secure Mainte nance solutions projects.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jeroen-keiren-model-check ing-in-the-context-of-digital-twins/ END:VEVENT BEGIN:VEVENT UID:20230502T1112Z-1683025941.8224-EO-1352-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230502T110102Z LAST-MODIFIED:20230502T110102Z DTSTART;TZID=Europe/Amsterdam:20230504T124500 DTEND;TZID=Europe/Amsterdam:20230504T133000 SUMMARY: Erik de Vink: In search of stability: a probabilistic composition of stable processes is stable DESCRIPTION: Axiomatization of branching bisimulation for distributions has been established in the setting of a process language with nondeterministi c and probabilistic choice in earlier work. Key ingredient to the proof of completeness of the proposed theory is the notion of stability of processes and the property that every distribution can evolve\, within the same equi valence class ...continue reading X-ALT-DESC;FMTTYPE=text/html:Axiomatization of branching bisimulation f or distributions has been established in the setting of a process language with nondeterministic and probabilistic choice in earlier work. Key ingredi ent to the proof of completeness of the proposed theory is the notion of st ability of
processes and the property that every distribution can evo lve\, within the same equivalence class modulo branching bisimilarity\, int o a stable distribution. In this talk we focus on a combinatorial approach to this result\, replacing the topological machinery used so far.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/erik-de-vink-in-search-of -stability-a-probabilistic-composition-of-stable-processes-is-stable/ END:VEVENT BEGIN:VEVENT UID:20230418T1129Z-1681817386.628-EO-1348-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230418T112751Z LAST-MODIFIED:20230418T112751Z DTSTART;TZID=Europe/Amsterdam:20230420T124500 DTEND;TZID=Europe/Amsterdam:20230420T133000 SUMMARY: Clemens Dubslaff: A hybrid modeling approach for feature-oriented systems DESCRIPTION: In feature-oriented software development\, software product li nes are defined as families of systems where features encapsulate increment al or optional behaviors. There are mainly two different approaches for spe cifying feature-oriented systems: either annotative of compositional\, i.e. \, either guarding code by presence conditions over features or specifying components for each feature along with a composition operation. Both ...con tinue reading X-ALT-DESC;FMTTYPE=text/html:In feature-oriented software development\, software product lines are defined as families of systems where features e ncapsulate incremental or optional behaviors. There are mainly two differen t approaches for specifying feature-oriented systems: either annotative of compositional\, i.e.\, either guarding code by presence conditions over fea tures or specifying components for each feature along with a composition op eration. Both approaches have their merits and downsides with respect to\, among others\, separation of concerns\, granularity\, extensibility\, and a nalyzability. One challenge posed by Apel and Kästner already in 2008 is to establish a hybrid approach that is both annotative and compositional\, un iting the benefits of both approaches. In this talk\, I will face this chal lenge from a formal perspective and illustrate how such a hybrid approach c an be achieved for feature-oriented systems specified in a variant of Dijks tra’s guarded command language. I will present extensions\, further benefit s for formal analysis of feature-oriented systems\, and how the results can be transferred back to the field of feature-oriented software development.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/clemens-dubslaff-a-hybrid -modeling-approach-for-feature-oriented-systems/ END:VEVENT BEGIN:VEVENT UID:20230411T0838Z-1681202299.0413-EO-1347-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230411T082251Z LAST-MODIFIED:20230411T082251Z DTSTART;TZID=Europe/Amsterdam:20230413T124500 DTEND;TZID=Europe/Amsterdam:20230413T133000 SUMMARY: Mark Bouwman: Algorithms for Cheaper LEGO DESCRIPTION: Besides the official LEGO sets there are many fan-created desi gns online. To build these custom designs you need to buy individual parts online. Concentrated on two larger platforms there are thousands of stores offering parts. An interesting optimisation problem arises: which distribut ion of parts over the stores is the cheapest\, also considering shipping co sts? In ...continue reading X-ALT-DESC;FMTTYPE=text/html:Besides the official LEGO sets there are m any fan-created designs online. To build these custom designs you need to b uy individual parts online. Concentrated on two larger platforms there are thousands of stores offering parts. An interesting optimisation problem ari ses: which distribution of parts over the stores is the cheapest\, also con sidering shipping costs? In the literature this is known as the Internet Sh opping Optimisation Problem (ISOP). In this talk I will present ISOP\, a pr oof that it is NP-Hard and a range of algorithms for obtaining (near) optim al solutions. Additionally\, I will present my own algorithm\, which combin es a number of techniques\, and show that it is effective in finding cheap stores for buying LEGO.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/mark-bouwman-algorithms-f or-cheaper-lego/ END:VEVENT BEGIN:VEVENT UID:20230404T1622Z-1680625357.9893-EO-1344-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230404T161002Z LAST-MODIFIED:20230404T161002Z DTSTART;TZID=Europe/Amsterdam:20230406T124500 DTEND;TZID=Europe/Amsterdam:20230406T133000 SUMMARY: Olav Bunte: An evaluation of the Spoofax language workbench in an industrial context DESCRIPTION: Language workbenches have been developed to ease the implement ation of Domain Specific Languages (DSLs)\, which are used often in industr y for better productivity and communication among engineers. However\, only very little literature exists that evaluates the use of a language workben ch for an industrial context. In a work in progress journal paper we do exa ctly ...continue reading X-ALT-DESC;FMTTYPE=text/html:Language workbenches have been developed t o ease the implementation of Domain Specific Languages (DSLs)\, which are u sed often in industry for better productivity and communication among engin eers. However\, only very little literature exists that evaluates the use o f a language workbench for an industrial context. In a work in progress jou rnal paper we do exactly this\, specifically for the language workbench Spo ofax. We use OIL\, developed at Canon Production Printing\, as the use case language that has been implemented in Spoofax over the past five years. In this presentation I focus on the evaluation of Spoofax on two aspects: pro ductivity (``How efficient is it to develop an industrial DSL with Spoofax compared to with a general purpose language?'') and modularity (``How modul ar can a language be implemented in Spoofax?'').
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/olav-bunte-an-evaluation- of-the-spoofax-language-workbench-in-an-industrial-context/ END:VEVENT BEGIN:VEVENT UID:20230328T0921Z-1679995314.657-EO-1343-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230328T091655Z LAST-MODIFIED:20230328T091655Z DTSTART;TZID=Europe/Amsterdam:20230330T124500 DTEND;TZID=Europe/Amsterdam:20230330T133000 SUMMARY: Ferry Timmers: Raking and the Self-deducing machine DESCRIPTION: Raking is a method with which we can extract the behavior from software systems by instrumenting its source code. A key element of this m ethod is something I have dubbed the self-deducing machine. In this talk I will go into more detail about this concept\, and how it can be used to ext ract a state ...continue reading X-ALT-DESC;FMTTYPE=text/html:Raking is a method with which we can extra ct the behavior from software systems by instrumenting its source code. A k ey element of this method is something I have dubbed the self-deducing mach ine. In this talk I will go into more detail about this concept\, and how i t can be used to extract a state space from an instrumented program by exec ution.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/ferry-timmers-raking-and- the-self-deducing-machine/ END:VEVENT BEGIN:VEVENT UID:20230320T1451Z-1679323862.3224-EO-1341-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230320T144821Z LAST-MODIFIED:20230320T144821Z DTSTART;TZID=Europe/Amsterdam:20230323T124500 DTEND;TZID=Europe/Amsterdam:20230323T133000 SUMMARY: Flip van Spaendonck: Verification of the busy-forbidden protocol u sing an extension of the cones and foci proof framework DESCRIPTION: In a previous colloquium\, we presented a new and efficient re aders-writer lock with no resource contention between readers\, called the Busy-Forbidden Protocol. For its verification\, specifications of its imple mentation and its less complex external behavior are provided. However\, we are unable to prove the equivalence of these models for more than 7 concur rent threads using ...continue reading X-ALT-DESC;FMTTYPE=text/html:In a previous colloquium\, we presented a new and efficient readers-writer lock with no resource contention between r eaders\, called the Busy-Forbidden Protocol.
For its verification\, s pecifications of its implementation and its less complex external behavior are provided.
However\, we are unable to prove the equivalence of the se models for more than 7 concurrent threads using mCRL2\, due to the state spaces becoming too large to work with.
We now give a general equival ence proof using our new extension of the cones and foci proof framework fo r divergence-preserving branching bisimilarity.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/flip-van-spaendonck-verif ication-of-the-busy-forbidden-protocol-using-an-extension-of-the-cones-and- foci-proof-framework/ END:VEVENT BEGIN:VEVENT UID:20230314T1058Z-1678791481.4826-EO-1335-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230314T104337Z LAST-MODIFIED:20230314T104337Z DTSTART;TZID=Europe/Amsterdam:20230316T124500 DTEND;TZID=Europe/Amsterdam:20230316T133000 SUMMARY: Jan Friso Groote: Real equation systems DESCRIPTION: The toolset heavily relies on boolean equation systems\, which are the workhorse to solve modal formulas. Boolean equation systems only p ermit truth values as solutions for variables. It would be nice to also ext ract quantitative information using modal formulas\, such as probabilities\ , durations\, or yields. For this purpose it would be nice to have real ... continue reading X-ALT-DESC;FMTTYPE=text/html:The toolset heavily relies on boolean equa tion systems\, which are the workhorse
to solve modal formulas. Boole an equation systems only permit truth values as solutions for
variabl es. It would be nice to also extract quantitative information using modal f ormulas\,
such as probabilities\, durations\, or yields. For this pur pose it would be nice to have
real equation systems\, i.e. equation s ystems where the fixed point variables represent
reals. In this talk I present how I look at them\, and show a systematic way to solve
the m manually. The work being presented is far from finished.
CATEGORIES:Colloquium LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-friso-groote-real-equ ation-systems/ END:VEVENT BEGIN:VEVENT UID:20230307T1251Z-1678193501.762-EO-1326-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230307T124559Z LAST-MODIFIED:20230307T124559Z DTSTART;TZID=Europe/Amsterdam:20230309T124500 DTEND;TZID=Europe/Amsterdam:20230309T133000 SUMMARY: Hans Zantema: Turtle figures of morphic sequences DESCRIPTION: In September I retired and as a present I received from you a great art object in stained glass. This was a particular instance of a turt le figure of a morphic sequence\, as will be described in this talk. The ma in message is that a program consisting of only a few lines may create a .. .continue reading X-ALT-DESC;FMTTYPE=text/html:In September I retired and as a present I received from you a great art object in stained glass.
This was a par ticular instance of a turtle figure of a morphic sequence\, as will be desc ribed in this talk.
The main message is that a program consisting of only a few lines may create a wide variety of figures\, some of which havin g amazing patterns. In 2016 my paper on this topic appeared in the journal Fractals. Recently I extended this work to a book in Dutch for a wider audi ence. This book just appeared\, and the title is ‘Spelen met oneindigheid\, verrassende figuren en patronen’\, see https://www.win.tue.nl/~hzantema/aank.pdf
In the talk some basic theory will be shown\, in particular leading to the staine d glass object.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/hans-zantema-turtle-figur es-of-morphic-sequences/ END:VEVENT BEGIN:VEVENT UID:20230227T1029Z-1677493740.1944-EO-1323-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230227T102819Z LAST-MODIFIED:20230227T102819Z DTSTART;TZID=Europe/Amsterdam:20230302T124500 DTEND;TZID=Europe/Amsterdam:20230302T133000 SUMMARY: Allan van Hulst: Kernels and small quasi-kernels in directed graph s (guest speaker) DESCRIPTION: In this talk I would like to discuss recent developments on th e subject of kernels and quasi-kernels in directed graphs. A kernel is an i ndependent set K in a directed graph such that every vertex is reachable in at most one step from K. A quasi-kernel is a weakening of the concept of a kernel. ...continue reading X-ALT-DESC;FMTTYPE=text/html:In this talk I would like to discuss recen t developments on the subject of kernels and quasi-kernels in directed grap hs.
A kernel is an independent set K in a directed graph such that ev ery vertex is reachable in at most one step from K. A quasi-kernel is a
weakening of the concept of a kernel. A quasi-kernel Q is an independent set in a directed graph such that every vertex can be
reached in at most two steps from Q. While there are simple directed graphs that do not h ave a kernel (for instance\, a directed
triangle) it is a well-known result that every directed graph must have at least one quasi-kernel. A sou rce in a directed graph is
a vertex having only outgoing arcs. It is an open problem since 1976 whether every source-free directed graph has a q uasi-kernel
of size at most half of the number of vertices. As shown in the preprint: https://arxiv.org/abs/2110.00789\, I h ave developed a
method to prove that the existence of a kernel implie s the existence of a 'small' quasi-kernel. In further work\, as shown in
https://arxiv.org/abs/2212.12764 I was able to prov e that\, if the number of sources introduced by the removal of a vertex and
its out-neighborhood can be bounded\, every source-free directed gra ph has a quasi-kernel of size at most the desired bound.
The latter m ethod uses a constructive approach that leads to an algorithm for finding a 'small' quasi-kernel
CATEGORIES:Colloquium LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/allan-van-hulst-kernels-a nd-small-quasi-kernels-in-directed-graphs-guest-speaker/ END:VEVENT BEGIN:VEVENT UID:20230222T1215Z-1677068103.0743-EO-1319-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230222T120831Z LAST-MODIFIED:20230222T120900Z DTSTART;TZID=Europe/Amsterdam:20230223T124500 DTEND;TZID=Europe/Amsterdam:20230223T133000 SUMMARY: Thomas Neele: Operations on Fixpoint Equation Systems\, a formalis ation in Coq. DESCRIPTION: Fixpoint equation systems over arbitrary complete lattices gen eralise several established formalisms such as parity games and parameteris ed Boolean equation systems. We identify a number of elementary operations\ , such as swapping equations and substituting variables by their definition s and prove their correctness. All proofs are formalised in Coq. In this ta lk\, I will go into the ...continue reading X-ALT-DESC;FMTTYPE=text/html:Fixpoint equation systems over arbitrary c omplete lattices generalise several established formalisms such as parity g ames and parameterised Boolean equation systems. We identify a number of el ementary operations\, such as swapping equations and substituting variables by their definitions and prove their correctness. All proofs are formalise d in Coq. In this talk\, I will go into the history of this work\, introduc e fixpoint equation systems\, show how these can be formalised in Coq\, dem o a couple of Coq proofs\, and highlight several Coq features.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/thomas-neele-restricting- sysml-activity-diagrams-to-facilitate-formal-analysis/ END:VEVENT BEGIN:VEVENT UID:20230214T1727Z-1676395660.4675-EO-1311-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230214T172011Z LAST-MODIFIED:20230214T172011Z DTSTART;TZID=Europe/Amsterdam:20230216T124500 DTEND;TZID=Europe/Amsterdam:20230216T133000 SUMMARY: Kevin Jilissen: Restricting SysML Activity Diagrams to facilitate formal analysis DESCRIPTION: SysML has several modeling techniques for defining the behavio r of a system. In this talk\, we have a closer look on modeling system beha vior using activity diagrams and the underlying semantics of these diagrams . Next\, we take a dive into the usage of SysML Activity Diagrams within t he generic tunnel control system models of Rijkswaterstaat. ...continue rea ding X-ALT-DESC;FMTTYPE=text/html:SysML has several modeling techniques for defining the behavior of a system. In this talk\, we have a closer look on modeling system behavior using activity diagrams and the underlying semanti cs of these diagrams. Next\, we take a dive into the usage of SysML Activi ty Diagrams within the generic tunnel control system models of Rijkswaterst aat. Inspired by this usage\, restrictions on the modeling language and its elements are defined which allow for an optimized translation to formal mo dels.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/kevin-jilissen-restrictin g-sysml-activity-diagrams-to-facilitate-formal-analysis/ END:VEVENT BEGIN:VEVENT UID:20230206T1217Z-1675685826.1599-EO-1310-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230206T121639Z LAST-MODIFIED:20230206T121639Z DTSTART;TZID=Europe/Amsterdam:20230209T124500 DTEND;TZID=Europe/Amsterdam:20230209T133000 SUMMARY: Rick Erkens: Optimizing Term Rewriting with Creeper Trace Transduc ers DESCRIPTION: In a previous talk I discussed our work on algorithms for opti mized term normalization. We use a so-called set automaton to find pattern matches efficiently\, and upon discovering a reducible subterm we apply the reduction while preserving as much matching information as possible. By in tegrating the structure of the right-hand sides into the automaton\, we ... continue reading X-ALT-DESC;FMTTYPE=text/html:In a previous talk I discussed our work on algorithms for optimized term normalization. We use a so-called set automa ton to find pattern matches efficiently\, and upon discovering a reducible subterm we apply the reduction while preserving as much matching informatio n as possible. By integrating the structure of the right-hand sides into th e automaton\, we can discover a new reducible subterm immediately\, and ext ract a trace of rewrite steps that can be executed in succession without ac tually executing the steps. In this talk I will discuss a way to exploit th is free information by focusing on a subset of traces. From a rewrite syste m we construct a so-called creeper trace transducer\, which reads a creeper trace while producing the term obtained after executing the steps in the t race. The transducer skips overlapping symbols between each pair of subsequ ent rules\, and in some cases a part of the trace can be disregarded altoge ther.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rick-erkens-optimizing-te rm-rewriting-with-creeper-trace-transducers/ END:VEVENT BEGIN:VEVENT UID:20230130T1026Z-1675074388.1828-EO-1309-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230130T102618Z LAST-MODIFIED:20230130T102618Z DTSTART;TZID=Europe/Amsterdam:20230202T124500 DTEND;TZID=Europe/Amsterdam:20230202T133000 SUMMARY: Herman Geuvers: Apartness and Hennessy-Milner logic DESCRIPTION: Apartness is the opposite (dual) of bisimulation. Intuitively\ , two states in a system are apart if there is a positive way to distinguis h them. Apartness is an inductive notion\, so we have a deduction system fo r proving that two states are apart\, and if we cannot prove they are apart \, they are bisimilar. This works ...continue reading X-ALT-DESC;FMTTYPE=text/html:Apartness is the opposite (dual) of bisimu lation. Intuitively\, two states in a system are apart if there is a positi ve way to distinguish them. Apartness is an inductive notion\, so we have a deduction system for proving that two states are apart\, and if we cannot prove they are apart\, they are bisimilar. This works for various notions o f bisimilarity\, especially for those where the systems can be described as co-algebras. So apartness gives an inductive view on the co-inductive noti on of bisimulation.
There is also a logical view on bisimulation: two states are bisimilar if-and-only-if they satisfy the same formulas of Henn essy-Milner logic (HML)\, where the precise syntax for HML formulas and the notion of satisfaction depends on the type of bisimulation one wants to ta lk about.
In the talk we will focus on this "if-and-only-if" in its d ual form\, using the inductive nature of apartness. We will prove (directly \, without referring to bisimulation): Two states are apart if-and-only-if there is a Hennessy-Milner logic formula that distinguishes them.
The "only if" is proved by constructing the HML formula by induction on the pr oof that two states are apart. The "if" part is proven by induction on the HML formula. We will discuss this for Labelled Transition Systems (LTS) wit h strong bisimulation\, weak bisimulation and branching bisimulation. For b ranching bisimulation\, the proof is remarkably tricky and we propose a sli ghtly different variant of HML\, PHMLU\, (Positive Hennessy-Milner Logic wi th Until) for branching bisimulation that simplifies the proof. We believe that PHMLU has merit of its own\, as being the "natural logic" for branchin g bisimulation/apartness.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/herman-geuvers-apartness- and-hennessy-milner-logic/ END:VEVENT BEGIN:VEVENT UID:20230124T1349Z-1674568167.9563-EO-1306-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230124T134137Z LAST-MODIFIED:20230124T134137Z DTSTART;TZID=Europe/Amsterdam:20230126T124500 DTEND;TZID=Europe/Amsterdam:20230126T133000 SUMMARY: Maurice Laveaux: Modern (Thread-Safe) Decision Diagram Library DESCRIPTION: Some time ago we have introduced a thread-safe term library in the mCRL2 toolset and have transformed some of the algorithms to paralleli sed variants. Similarly\, we rely on a thread-safe decision diagram with pa rallelised algorithms for computing symbolic reachability and bisimulation. These (highly efficient) implementations often sacrifice readability of so urce code for the purpose of ...continue reading X-ALT-DESC;FMTTYPE=text/html:Some time ago we have introduced a thread- safe term library in the mCRL2 toolset and have transformed some of the alg orithms to parallelised variants. Similarly\, we rely on a thread-safe deci sion diagram with parallelised algorithms for computing symbolic reachabili ty and bisimulation. These (highly efficient) implementations often sacrifi ce readability of source code for the purpose of achieving performance\, wh ich makes it hard to extend them and do experiments. Furthermore\, both lib raries make different design choices in their implementation. In this talk I will be evaluating the design decisions that have been made and give insi ghts into possible alternatives based on a prototype implementation.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/maurice-laveaux-modern-th read-safe-decision-diagram-library/ END:VEVENT BEGIN:VEVENT UID:20230116T1032Z-1673865177.5435-EO-1297-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230116T103129Z LAST-MODIFIED:20230116T103129Z DTSTART;TZID=Europe/Amsterdam:20230119T124500 DTEND;TZID=Europe/Amsterdam:20230119T133000 SUMMARY: Jeroen Keiren: An alternative characterization of fixed points DESCRIPTION: To establish that some set X is a subset of greatest fixpoint nu f\, we routinely show that X is a post-fixpoint\, ie.\, X is a subset of f(X). This is a straightforward\, constructive way of proving greatest fix points. However\, for least fixpoints\, constructive approaches such as fix point iteration break down when we consider subset ...continue reading X-ALT-DESC;FMTTYPE=text/html:To establish that some set X is a subset o f greatest fixpoint nu f\, we routinely show that X is a post-fixpoint\, ie .\, X is a subset of f(X). This is a straightforward\, constructive way of proving greatest fixpoints. However\, for least fixpoints\, constructive ap proaches such as fixpoint iteration break down when we consider subset latt ices over infinite sets.
In this talk\, we introduce support ordering
s\, along with general lattice-theoretical results\, that formalize the dep
endencies between states that satisfy given fixpoint formulas. This\, in es
sence\, gives a purely semantic\, constructive account of the least and gre
atest fixpoints over monotonic functions over subset lattices. We will show
how this can be used to reason about least fixpoints in the mu-calculus ov
er infinite-state systems.
This is joint work with Rance Cleaveland.
p>
CATEGORIES:Colloquium
LOCATION:MF 6.132
GEO:51.447551;5.487453
ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl
URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jeroen-keiren-an-alternat
ive-characterization-of-fixed-points/
END:VEVENT
BEGIN:VEVENT
UID:20230109T1007Z-1673258862.9086-EO-1295-1@37.128.148.44
STATUS:CONFIRMED
DTSTAMP:20240319T075845Z
CREATED:20230109T095618Z
LAST-MODIFIED:20230109T120434Z
DTSTART;TZID=Europe/Amsterdam:20230112T124500
DTEND;TZID=Europe/Amsterdam:20230112T133000
SUMMARY: Jan Martens: Computing minimal distinguishing HML-formulas is NP-h
ard
DESCRIPTION: The Hennessy-Milner Logic (HML) is a modal logic that expresse
s behavioural properties of states in LTSs. We are interested in explaining
behavioural inequivalence by constructing a formula that /distinguishes/ a
pair of states\, i.e. a formula that is true in exactly one of the states.
Cleaveland presented a method to generate a HML-formula by back-tracking .
..continue reading
X-ALT-DESC;FMTTYPE=text/html:
The Hennessy-Milner Logic (HML) is a modal logic that expresses behavioural properties of states in LTSs. We are inte rested in explaining behavioural inequivalence by constructing a formula th at /distinguishes/ a pair of states\, i.e. a formula that is true in exactl y one of the states. Cleaveland presented a method to generate a HML-formul a by back-tracking information from a partition refinement algorithm like t he Kanellakis-Smolka algorithm. This method yields a formula which is "oft en minimal in a precisely defined sense" - [1]. In this talk we cover the f ollowing:
2.) Sketch a polynom ial time algorithm to compute distinguishing formulas minimal with respect to observation-depth (number of nested diamond modalities) and negation-dep th (number of nested negations).
3.) Show by a reduction from CNF-SAT that the problem of computing a minimal distinguishing formula is NP-hard.
[1] Cleaveland\, Rance. "On automatically explaining bisimulation in equivalence." International Conference on Computer Aided Verification. Spri nger\, Berlin\, Heidelberg\, 1991.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-martens-computing-min imal-distinguishing-hml-formulas-is-np-hard/ END:VEVENT BEGIN:VEVENT UID:20230103T1207Z-1672747663.9071-EO-1294-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20230103T120330Z LAST-MODIFIED:20230103T120330Z DTSTART;TZID=Europe/Amsterdam:20230105T124500 DTEND;TZID=Europe/Amsterdam:20230105T133000 SUMMARY: Tim Willemse: On a relation between Failures Refinement and Model- Based Testing DESCRIPTION: Modelling the behaviour of a system\, and stepwise refining th at model until it has become sufficiently detailed to generate executable i mplementation is an appealing approach to software development\, advocated by companies such as\, e.g. Verum. The approach is rooted in solid mathemat ics and supported by tools\, such as mCRL2\, that help the developer. Under the ...continue reading X-ALT-DESC;FMTTYPE=text/html:Modelling the behaviour of a system\, and stepwise refining that model until it has become sufficiently detailed to g enerate executable implementation is an appealing approach to software deve lopment\, advocated by companies such as\, e.g. Verum. The approach is root ed in solid mathematics and supported by tools\, such as mCRL2\, that help the developer. Under the assumption that these tools are correct\, the rema ining point of attention in this approach is in correctly generating and ex ecuting the implementation from a model. Since it is very hard to reason fo rmally about the implementation and the platform on which it runs\, arguing that the generated software is ready to be shipped is tricky. In a standar d software development trajectory\, this is often decided by running test c ases against the implementation. In this talk\, I will entertain the idea o f testing the generated implementation using model-based testing\, exploiti ng the already existing models.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tim-willemse-on-a-relatio n-between-failures-refinement-and-model-based-testing/ END:VEVENT BEGIN:VEVENT UID:20221221T1013Z-1671617633.6688-EO-1290-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20221221T101053Z LAST-MODIFIED:20221221T101053Z DTSTART;TZID=Europe/Amsterdam:20221221T130000 DTEND;TZID=Europe/Amsterdam:20221221T141500 SUMMARY: Milan Hutten: Sound and Complete Axiomatizations for the rooted va riants of Divergence-Preserving\, Weak Divergence-Preserving and Stability- Respecting Branching Bisimilarity DESCRIPTION: This thesis studies divergence and the extra complexity it add s to branching bisimilarity and axiomatizations with respect to branching b isimilarity. Aceto et al. provided sound and complete axiomatization with r espect to rooted branching bisimilarity over basic CCS with the prefix iter ation operator. Additionally\, Spaninks\, and Liu and Yu proposed sound and complete axiomatizations with respect ...continue reading X-ALT-DESC;FMTTYPE=text/html:This thesis studies divergence and the ext ra complexity it adds to branching bisimilarity and axiomatizations with re spect to branching bisimilarity. Aceto et al. provided sound and complete a xiomatization with respect to rooted branching bisimilarity over basic CCS with the prefix iteration operator. Additionally\, Spaninks\, and Liu and Y u proposed sound and complete axiomatizations with respect to rooted diverg ence-preserving branching bisimilarity over basic CCS with the recursion co nstruct. We contribute by providing a sound and complete axiomatization wit h respect to rooted divergence-preserving branching bisimilarity over basic CCS with the prefix iteration operator. Additionally\, we provide sound an d complete axiomatizations with respect to rooted weakly divergence-preserv ing branching bisimilarity and rooted stability-respecting branching bisimi larity over the same algebra and show that completeness of these axiomatiza tions and the axiomatization with respect to rooted branching bisimilarity provided by Aceto et al. can be derived from the completeness of the axioma tization with respect to rooted divergence-preserving branching bisimilarit y. Lastly\, we discuss how the prefix iteration operator can be expressed t hrough the recursion construct and partially show how our proposed axiomati zation is represented in the axiomatizations of Spaninks\, and Liu and Yu.< /p> CATEGORIES:MSc Defence LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Bas Luttik":MAILTO:s.p.luttik@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/milan-hutten-sound-and-co mplete-axiomatizations-for-the-rooted-variants-of-divergence-preserving-wea k-divergence-preserving-and-stability-respecting-branching-bisimilarity/ END:VEVENT BEGIN:VEVENT UID:20221213T1648Z-1670950082.0475-EO-1288-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20221213T164657Z LAST-MODIFIED:20221213T164657Z DTSTART;TZID=Europe/Amsterdam:20221215T124500 DTEND;TZID=Europe/Amsterdam:20221215T133000 SUMMARY: Erik de Vink: Zielonka’s algorithm for variability parity games DESCRIPTION: Model checking approaches to software product lines are doubly cursed. Not only high-dimensionality regarding the number of states may be troublesome\, also the exponential number of options for a product raises a computational hindrance. In order to mitigate the latter\, family-based v erification\, as opposed to product-based verification\, has been proposed. In this talk we consider ...continue reading X-ALT-DESC;FMTTYPE=text/html:
Model checking approaches to software prod uct lines are doubly cursed. Not only high-dimensionality regarding the num ber of states may be troublesome\, also the exponential number of options f or a product raises a computational hindrance. In order to mitigate the lat ter\, family-based verification\, as opposed to product-based verification\ , has been proposed.
In this talk we consider so-called variability p arity games (VPGs). Given a set of possible options\, VPGs aggregate a corr esponding set of parity games. We discuss how the well-known recursive algo rithm to solve standard parity games attributed to Zielonka can be adapted to deal with VPGs. Experiments indicate that family-based verification base d on VPGs outperforms product-based verification based on parity games. In a search for an explanation of this we conjecture a relationship between th e number of recursive calls in Zielonka's algorithm for the case of a VPG v s. the parity games its consists of.
Joint work with Maurice ter Beek (CNR/ISTI Pisa) and Tim Willemse.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/erik-de-vink-zielonkas-al gorithm-for-variability-parity-games/ END:VEVENT BEGIN:VEVENT UID:20221205T1406Z-1670249173.387-EO-1278-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20221205T140523Z LAST-MODIFIED:20221205T140523Z DTSTART;TZID=Europe/Amsterdam:20221208T124500 DTEND;TZID=Europe/Amsterdam:20221208T133000 SUMMARY: Anna Stramaglia: lpsparunfold: features and application DESCRIPTION: Concise process models in mCRL2 can be obtained by the use of data types such as lists and structured sorts. However\, the addition of st ructure in the process parameters negatively affects static analysis tools such as constant elimination\, parameter elimination and sum elimination wh ich consider process parameters as single units in their analysis. To addre ss ...continue reading X-ALT-DESC;FMTTYPE=text/html:Concise process models in mCRL2 can be obt ained by the use of data types such as lists and structured sorts. However\ , the addition of structure in the process parameters negatively affects st atic analysis tools such as constant elimination\, parameter elimination an d sum elimination which consider process parameters as single units in thei r analysis. To address this problem\, Groote and Lisser introduced a techni que for flattening the structure of process parameters such that other stat ic analysis tools can apply their transformation more effectively. This tec hnique is implemented in the mCRL2 tool lpsparunfold. In practice\, we obse rve that\, although parameters can be unfolded\, the static analysis tools do not always benefit from this unfolding. On the other hand we observe tha t the unfolding seems quite beneficial for symbolic reachability (lpsreach) . In this talk I will introduce the lpsparunfold technique with a running e xample. I will talk about the relevant lpsparunfold options (with examples) and discuss the results of their application
CATEGORIES:Colloquium LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/anna-stramaglia-lpsparunf old-features-and-application/ END:VEVENT BEGIN:VEVENT UID:20221130T1348Z-1669816097.563-EO-1271-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20221130T134531Z LAST-MODIFIED:20221130T134531Z DTSTART;TZID=Europe/Amsterdam:20221201T124500 DTEND;TZID=Europe/Amsterdam:20221201T133000 SUMMARY: Tom Franken: Autonomous Data Language for Parallel Programming DESCRIPTION: In this presentation\, I shall introduce Autonomous Data Langu age\, highlight its most important attributes and explain the main concepts of the language using some examples. X-ALT-DESC;FMTTYPE=text/html:In this presentation\, I shall introduce A utonomous Data Language\, highlight its most important attributes and expla in the main concepts of the language using some examples.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tom-franken-autonomous-da ta-language-for-parallel-programming/ END:VEVENT BEGIN:VEVENT UID:20221121T1219Z-1669033145.5772-EO-1270-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20221121T121145Z LAST-MODIFIED:20221121T121145Z DTSTART;TZID=Europe/Amsterdam:20221124T124500 DTEND;TZID=Europe/Amsterdam:20221124T133000 SUMMARY: Valentina Castiglioni: Uncertainties\, adaptability\, and verifica tion DESCRIPTION: The principal objective of our ongoing IRF project “Programs i n the wild: Uncertainties\, adaptability\, and verification” is to provide a formal framework and tools for modelling and verifying the behaviour of s ystems characterised by a close interaction of a program with an unpredicta ble environment\, like Cyber-Physical Systems (CPSs). In this talk we will discuss the ...continue reading X-ALT-DESC;FMTTYPE=text/html:The principal objective of our ongoing IRF
project "Programs in the wild: Uncertainties\, adaptability\, and veri
fication" is to provide a formal framework and tools for modelling
and verifying the behaviour of systems characterised by a cl
ose interaction of a program with an unpredictable environment\, like Cyber
-Physical Systems (CPSs).
In this talk we will discuss the Robustn
ess Temporal Logic (RobTL)\, a novel temporal logic for the s
pecification and analysis of distances between the behaviours of C
PSs over a finite time horizon. Differently from classical temporal logic e
xpressing properties on the behaviour of a system\, we can use RobTL specif
ications to measure the differences in the behaviours of systems with respe
ct to various objectives and temporal constraints\, and to study how those
differences evolve in time.
As the behaviour of CPSs is inevitably su
bject to uncertainties and approximations\, we show how the unique features
of RobTL allow us to specify property of robustness of systems
On November 22\, 2022\, Maurice Laveaux wi ll defend his thesis titled ‘Accelerated Verification of Concurrent Systems '\, which is available via https://research.tue.nl/en/publications/accelerated-verification-of-co ncurrent-systems. The defence will also be streamed online via MS Teams . The link to the stream is available on request from m.laveaux@tue.nl.
CATEGORIES:PhD Defence LOCATION:Atlas 0.710 GEO:51.446751;5.486286 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/maurice-laveaux-accelerat ed-verification-of-concurrent-systems/ END:VEVENT BEGIN:VEVENT UID:20221115T0922Z-1668504163.3045-EO-1262-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20221115T091940Z LAST-MODIFIED:20221115T091940Z DTSTART;TZID=Europe/Amsterdam:20221117T124500 DTEND;TZID=Europe/Amsterdam:20221117T133000 SUMMARY: Clemens Dubslaff: “Feature Causality” DESCRIPTION: Almost all practical software systems today are configurable. Huge configuration spaces\, usually of size exponential in the number of co nfiguration options or features\, render their design\, analysis\, and expl anation challenging tasks. In this talk\, I will introduce the notion of “f eature causality” to support explainability of configurable systems. Inspir ed by the seminal definition of actual ...continue reading X-ALT-DESC;FMTTYPE=text/html:Almost all practical software systems toda y are configurable. Huge configuration spaces\, usually of size exponential in the number of configuration options or features\, render their design\, analysis\, and explanation challenging tasks. In this talk\, I will introd uce the notion of “feature causality” to support explainability of configur able systems. Inspired by the seminal definition of actual causality by Hal pern and Pearl\, feature causes capture configuration decisions that are re asons for fulfilling functional and non-functional system properties\, e.g. \, safety requirements or quality of service\, respectively. I will present various methods to explicate such reasons\, e.g.\, based on well-establish ed notions of responsibility and blame. By means of an evaluation on a wide range of configurable software systems\, including community benchmarks an d real-world systems\, I will demonstrate feasibility of this approach to i dentify root causes\, estimate the effects of configuration options\, and d etect feature interactions.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/clemens-dubslaff-feature- causality/ END:VEVENT BEGIN:VEVENT UID:20221109T0953Z-1667987622.5332-EO-1260-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20221109T093558Z LAST-MODIFIED:20221110T124325Z DTSTART;TZID=Europe/Amsterdam:20221110T124500 DTEND;TZID=Europe/Amsterdam:20221110T133000 SUMMARY: Bas Luttik: About divergence-preserving branching bisimilarity X-ALT-DESC;FMTTYPE=text/html: CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/bas-luttik-verifying-mutu al-exclusion-algorithms-dropping-the-atomicity-assumpti/ END:VEVENT BEGIN:VEVENT UID:20221103T1008Z-1667470094.437-EO-1259-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20221103T100011Z LAST-MODIFIED:20221103T100011Z DTSTART;TZID=Europe/Amsterdam:20221103T124500 DTEND;TZID=Europe/Amsterdam:20221103T133000 SUMMARY: Jan Friso Groote: Another attempt to make BDDs more compact. DESCRIPTION: Quite some time ago I presented an approach to represent BDDs more compactly. The approach could be used to represent some formulas expon entially more compact than in ordinary BDDs. However\, it was not possible to prove that this representation was never more than polynomially larger t han BDDs. This approach turned out to be a particular ...continue reading X-ALT-DESC;FMTTYPE=text/html:Quite some time ago I presented an approac h to represent BDDs more compactly. The approach could be used to represent some formulas exponentially more compact than in ordinary BDDs. However\, it was not possible to prove that this representation was never more than p olynomially larger than BDDs. This approach turned out to be a particular c ase of SDDs.
For some time I have been working to improve the approac h\, and I believe that this time I have an approach that works\, although q uite a number of aspects have not yet been proven. I will present what I ha ve up till now.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-friso-groote-another- attempt-to-make-bdds-more-compact/ END:VEVENT BEGIN:VEVENT UID:20221027T1154Z-1666871667.2549-EO-1251-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20221027T114115Z LAST-MODIFIED:20221027T114115Z DTSTART;TZID=Europe/Amsterdam:20221027T124500 DTEND;TZID=Europe/Amsterdam:20221027T133000 SUMMARY: Mark Bouwman: Decompositional Branching Bisimulation Minimisation of Monolithic Processes DESCRIPTION: A well known technique to reduce the impact of the state space explosion problem problem is compositional minimisation. In this technique \, first the state spaces of all components are computed and minimised modu lo some behavioural equivalence (e.g.\, some form of bisimilarity). These m inimised transition systems are subsequently combined to obtain the final s tate space. In ...continue reading X-ALT-DESC;FMTTYPE=text/html:A well known technique to reduce the impac t of the state space explosion problem problem is compositional minimisatio n. In this technique\, first the state spaces of all components are compute d and minimised modulo some behavioural equivalence (e.g.\, some form of bi similarity). These minimised transition systems are subsequently combined t o obtain the final state space.
In earlier work (by Maurice and Tim) a compositional minimisation technique was presented tailored to mCRL2: it provides support for the multi-action semantics of mCRL2 and allows splitti ng up a monolithic linear process specification into components. Only stron g bisimulation minimisation of components could be used\, limiting the effe ctiveness of the approach. In this talk I will present an extension to supp ort branching bisimulation reduction. Additionally\, I will present a numbe r of benchmarks using mCRL2 models derived from railway SysML models\, show ing that a significant reduction can be achieved\, also compared to composi tional minimisation with strong bisimulation reduction.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/mark-bouwman-decompositio nal-branching-bisimulation-minimisation-of-monolithic-processes/ END:VEVENT BEGIN:VEVENT UID:20221020T1438Z-1666276726.2708-EO-1245-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20221020T143754Z LAST-MODIFIED:20221020T143754Z DTSTART;TZID=Europe/Amsterdam:20221020T124500 DTEND;TZID=Europe/Amsterdam:20221020T133000 SUMMARY: Flip van Spaendonck: Efficient dynamic model based testing of mode ls with run-to-completion semantics DESCRIPTION: Model based testing (MBT) provides an efficient and automated approach to finding discrepancies between software models and their impleme ntation. If we want to incorporate MBT into the software development proces s\, i.e. code changes are only accepted if the implementation remains in co nformance with the model\, then MBT must be able to thoroughly test the ent ire ...continue reading X-ALT-DESC;FMTTYPE=text/html:Model based testing (MBT) provides an effi cient and automated approach to finding discrepancies between software mode ls and their implementation. If we want to incorporate MBT into the softwar e development process\, i.e. code changes are only accepted if the implemen tation remains in conformance with the model\, then MBT must be able to tho roughly test the entire model in as little time as possible. However\, curr ent academic MBT tools traverse models at-random. While this allows them to find most discrepancies relatively quickly\, it is inefficient at testing all or almost all states in the software model and is thus lacking in thoro ughness.
We provide a new traversal strategy and show that the at-ran dom strategy needs to take significantly longer to test >95 of all states. The benchmarks are performed on pseudo-randomly generated LTS's\, which we show to be representative of systems with run-to-completion semantics.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/flip-van-spaendonck-effic ient-dynamic-model-based-testing-of-models-with-run-to-completion-semantics / END:VEVENT BEGIN:VEVENT UID:20221013T1300Z-1665666058.934-EO-1242-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20221013T125149Z LAST-MODIFIED:20221013T125149Z DTSTART;TZID=Europe/Amsterdam:20221013T124500 DTEND;TZID=Europe/Amsterdam:20221013T133000 SUMMARY: Nobuko Yoshida: mCRL2 for type-based verifications for distributed processes and programs DESCRIPTION: I first summarise how we used mCRL2 for checking properties of distributed processes and programs. I then talk about our most recent work which applies mCRL2 to unreliable systems\, which we presented at CONCUR 2 022. X-ALT-DESC;FMTTYPE=text/html:I first summarise how we used mCRL2 for ch ecking properties of distributed processes and programs. I then talk about our most recent work which applies mCRL2 to unreliable systems\, which we p resented at CONCUR 2022.
CATEGORIES:Colloquium LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/nobuko-yoshida-mcrl2-for- type-based-verifications-for-distributed-processes-and-programs/ END:VEVENT BEGIN:VEVENT UID:20221003T1217Z-1664799468.8432-EO-1237-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20221003T121320Z LAST-MODIFIED:20221003T121320Z DTSTART;TZID=Europe/Amsterdam:20220929T124500 DTEND;TZID=Europe/Amsterdam:20220929T133000 SUMMARY: Olav Bunte: Implementing OIL in Spoofax DESCRIPTION: In previous presentations about OIL I have mainly dived into i ts semantics\, but in this presentation I’ll focus on its implementation in the Spoofax language workbench instead\, which has been ongoing work for 5 years now. I’ll discuss how we tackled different implementation aspects of a language\, such as syntax\, transformations and static semantics\, and e valuate ...continue reading X-ALT-DESC;FMTTYPE=text/html:In previous presentations about OIL I have mainly dived into its semantics\, but in this presentation I'll focus on i ts implementation in the Spoofax language workbench instead\, which has bee n ongoing work for 5 years now. I'll discuss how we tackled different imple mentation aspects of a language\, such as syntax\, transformations and stat ic semantics\, and evaluate Spoofax in how well it supported us in this.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/olav-bunte-implementing-o il-in-spoofax/ END:VEVENT BEGIN:VEVENT UID:20220927T1341Z-1664286096.0968-EO-1232-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220927T133110Z LAST-MODIFIED:20220927T133110Z DTSTART;TZID=Europe/Amsterdam:20220922T124500 DTEND;TZID=Europe/Amsterdam:20220922T133000 SUMMARY: Anton Wijs: GPU hash tables DESCRIPTION: Hash tables are important data structures for the fast storage and retrieval of data elements. They are\, for instance\, used often in ex plicit-state model checkers. In this talk I take a closer look at different types of hash tables\, and address their suitability for use on graphics p rocessors (GPUs). The current version of GPUexplore\, an ...continue readin g X-ALT-DESC;FMTTYPE=text/html:Hash tables are important data structures for the fast storage and retrieval of data elements. They are\, for instanc e\, used often in explicit-state model checkers. In this talk I take a clos er look at different types of hash tables\, and address their suitability f or use on graphics processors (GPUs). The current version of GPUexplore\, a n explicit-state model checker running entirely on a GPU\, contains impleme ntations of various combinations of hashing techniques.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/anton-wijs-gpu-hash-table s/ END:VEVENT BEGIN:VEVENT UID:20220919T1435Z-1663598142.5471-EO-1226-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220919T142414Z LAST-MODIFIED:20220919T142414Z DTSTART;TZID=Europe/Amsterdam:20220908T124500 DTEND;TZID=Europe/Amsterdam:20220908T133000 SUMMARY: Ferry Timmers: Formalising Raking DESCRIPTION: Raking is a technique I have developed that allows extracting a statespace of software systems in situ. It does not query the system\, bu t instead uses annotation of the system’s program code. In this talk\, I wi ll go deeper into the semantics of this technique. X-ALT-DESC;FMTTYPE=text/html:Raking is a technique I have developed that allow s extracting a statespace of software systems in situ. It does not query th e system\, but instead uses annotation of the system's program code. In thi s talk\, I will go deeper into the semantics of this technique.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/ferry-timmers-formalising -raking/ END:VEVENT BEGIN:VEVENT UID:20220822T0730Z-1661153427.3612-EO-1189-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220822T072431Z LAST-MODIFIED:20220829T123918Z DTSTART;TZID=Europe/Amsterdam:20220822T150000 DTEND;TZID=Europe/Amsterdam:20220822T154500 SUMMARY: Kevin Jilissen: A formal analysis of the tunnel control systems of the Rijkswaterstaat GITO DESCRIPTION: On Monday\, August 29\, at 15:00 in MF13 Kevin Jilissen defend s his master thesis called A formal analysis of the tunnel control systems of the Rijkswaterstaat GITO. X-ALT-DESC;FMTTYPE=text/html:On Monday\, August 29\, at 15:00 in MF13 K evin Jilissen defends his master thesis called A formal analysis of the tunnel control systems of the Rijkswaterstaat GITO.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 13 GEO:51.447532;5.487437 ORGANIZER;CN="Jan Friso Groote":MAILTO:j.f.groote@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/kevin-jilissen-graduates- on-tunnel-control-systems/ END:VEVENT BEGIN:VEVENT UID:20220714T0910Z-1657789820.0809-EO-1178-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220714T082925Z LAST-MODIFIED:20220714T082925Z DTSTART;TZID=Europe/Amsterdam:20220714T124500 DTEND;TZID=Europe/Amsterdam:20220714T133000 SUMMARY: Rance Cleaveland: QUERY-CHECKING FOR FINITE LINEAR-TIME TEMPORAL L OGIC DESCRIPTION: This talk addresses the following problem: given a finite set of observations of system behavior\, each observation itself being a finit e sequence of system states\, and a query consisting of a Finite LTL formul a with missing subformulas\, compute missing the missing subformulas that m ake the LTL formula true for all observations. This so-called query-checki ng problem ...continue reading X-ALT-DESC;FMTTYPE=text/html:
This talk addresses the following problem: given a finite set of observations of system behavior\, each observation itself being a finite sequence of system states\, and a query consisting of a Finite LTL formula with missing subformulas\, compute missing the missin g subformulas that make the LTL formula true for all observations. This so -called query-checking problem has many applications in the analysis of tim e-series data\, including server logs and financial trend data\, as well as generally in system comprehension and system-specification mining. The pr esentation will begin by introducing Finite LTL\, which is the well-known L inear-Time Temporal Logic of Pnueli interpreted over finite\, rather than i nfinite\, state sequences. It will then show how automata may be construct ed from such formulas that recognize exactly the sequences making the formu la true\, and how these automata may be adapted to Finite LTL queries in wh ich subformulas are missing. A procedure is then described that uses the a utomaton representation of such a query to solve for the missing subformula s for a given a set of finite state sequences. Preliminary experimental re sults of a prototype implementation are also given.
CATEGORIES:Colloquium LOCATION:MetaForum MF 14 GEO:51.447532;5.487437 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rance-cleaveland-query-ch ecking-for-finite-linear-time-temporal-logic/ END:VEVENT BEGIN:VEVENT UID:20220705T1428Z-1657031288.9703-EO-1174-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220705T082025Z LAST-MODIFIED:20220705T082025Z DTSTART;TZID=Europe/Amsterdam:20220707T124500 DTEND;TZID=Europe/Amsterdam:20220707T133000 SUMMARY: Tim Willemse: On Model-Based Testing DESCRIPTION: Model-Based Testing is a formal approach to testing (software) systems. Input-output conformance\, often abbreviated by “ioco”\, is one s uch approach to formal testing. It assumes that implementations can be mode lled by input-enabled input-output transition systems\, and it formalises w hen a given implementation conforms to a specification of that implementati on. In the context of TNO-related projects\, ...continue reading X-ALT-DESC;FMTTYPE=text/html:Model-Based Testing is a formal approach t o testing (software) systems. Input-output conformance\, often abbreviated by "ioco"\, is one such approach to formal testing. It assumes that impleme ntations can be modelled by input-enabled input-output transition systems\, and it formalises when a given implementation conforms to a specification of that implementation. In the context of TNO-related projects\, we have be en studying (1) the problem of using ioco-based testing in the setting of V erum\, and (2) the problem of deriving 'difference' specifications that are to be used for minimising the ioco-based testing effort. In this talk\, I will explain the basic philosophy and theory behind ioco\, discuss some of its peculiarities\, and\, (preparation) time permitting\, I hope to address at least one of the two mentioned TNO-related theoretical problems.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tim-willemse-on-model-bas ed-testing/ END:VEVENT BEGIN:VEVENT UID:20220628T1129Z-1656415753.4512-EO-1170-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220628T110952Z LAST-MODIFIED:20220628T110952Z DTSTART;TZID=Europe/Amsterdam:20220630T124500 DTEND;TZID=Europe/Amsterdam:20220630T133000 SUMMARY: Maurice Laveaux: Signature-based symbolic bisimulation minimisatio n DESCRIPTION: Symbolic exploration techniques allow us to construct symbolic state spaces with billions of states. Reducing these state spaces modulo ( strong) bisimulation would help in further analysis. In this talk I will pr esent a signature-based partition refinement algorithm from the literature to compute (strong) bisimulation on a symbolic state space. This algorithm requires a transformation of ...continue reading X-ALT-DESC;FMTTYPE=text/html:Symbolic exploration techniques allow us t o construct symbolic state spaces with billions of states. Reducing these s tate spaces modulo (strong) bisimulation would help in further analysis. In this talk I will present a signature-based partition refinement algorithm from the literature to compute (strong) bisimulation on a symbolic state sp ace. This algorithm requires a transformation of the symbolic state space t hat we would like to avoid. Therefore\, I will also discuss the challenges in adapting this algorithm to our setting.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/maurice-laveaux-signature -based-symbolic-bisimulation-minimisation/ END:VEVENT BEGIN:VEVENT UID:20220617T0532Z-1655443965.0388-EO-1162-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220616T182602Z LAST-MODIFIED:20220616T182602Z DTSTART;TZID=Europe/Amsterdam:20220623T150000 DTEND;TZID=Europe/Amsterdam:20220623T154500 SUMMARY: Floris Zeven: Spatial Model Checking with mCRL2 DESCRIPTION: Image analysis using spatial model checking is a relatively re cent approach that has promising applications in the medical field. We inve stigated whether it is possible to verify spatial proper- ties using the mC RL2 toolset\, which was built to verify concurrent systems and protocols. T his was achieved by translating Spatial Logic for Closure Spaces (SLCS) for mulae ...continue reading X-ALT-DESC;FMTTYPE=text/html:Image analysis using spatial model checkin g is a relatively recent approach that has promising applications in the me dical field. We investigated whether it is possible to verify spatial prope r- ties using the mCRL2 toolset\, which was built to verify concurrent syst ems and protocols. This was achieved by translating Spatial Logic for Closu re Spaces (SLCS) formulae to μ-calculus formulae\, proving this translation is correct\, and by creating a script that associates an mCRL2 specificati on with an image. As mCRL2 only verifies properties on an initial state\, w hich would correspond to a single pixel\, a proof-of-concept implementation was subsequently created that utilizes the mCRL2 toolset to mark every pix el that satisfies a spatial property.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Bas Luttik":MAILTO:s.p.luttik@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/floris-zeven-spatial-mode l-checking-with-mcrl2/ END:VEVENT BEGIN:VEVENT UID:20220623T1158Z-1655985513.3565-EO-1169-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220623T115027Z LAST-MODIFIED:20220623T115027Z DTSTART;TZID=Europe/Amsterdam:20220623T124500 DTEND;TZID=Europe/Amsterdam:20220623T133000 SUMMARY: Anna Stramaglia: A journey across Cordis models and their verifica tion DESCRIPTION: In the MACHINAIDE project we verify Cordis models by means of mCRL2. Cordis models are UML-like models developed\, tested and simulated i n the Cordis SUITE. In this presentation I will take you along on our journ ey towards verification of Cordis models in three parts: Description of the structure and semantics of Cordis models Peculiarities in the ...continue reading X-ALT-DESC;FMTTYPE=text/html:In the MACHINAIDE project we verify Cordis models by means of mCRL2. Cordis models are UML-like models developed\, te sted and simulated in the Cordis SUITE.
It is a natural question to find a DFA or NFA for which a given set of words should be accepted and another given set should not be accepted.
In this presentation we investigate how to f ind a smallest automaton for both types by means of SMT solving\, and compa re the results.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/hans-zantema-passive-auto mata-learning-dfas-and-nfas/ END:VEVENT BEGIN:VEVENT UID:20220607T1403Z-1654610610.7726-EO-1158-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220607T115347Z LAST-MODIFIED:20220607T115606Z DTSTART;TZID=Europe/Amsterdam:20220609T124500 DTEND;TZID=Europe/Amsterdam:20220609T133000 SUMMARY: Jeroen Keiren: On the semantics of data types in mCRL2 DESCRIPTION: In my previous talk I described the unfolding of process param eters in mCRL2\, as it is done by lpsparunfold. This technique requires ext ending data specifications with new operations and equations. As part of th e correctness\, we need to reason about properties of the data specificatio n. Working on these proofs triggered questions about the mCRL2 data ...cont inue reading X-ALT-DESC;FMTTYPE=text/html:In my previous talk I described the unfold ing of process parameters in mCRL2\, as it is done by lpsparunfold.
T his technique requires extending data specifications with new operations an d equations. As part of the correctness\, we need to reason about propertie s of the data specification. Working on these proofs triggered questions ab out the mCRL2 data types.
In this talk\, I will discuss the progress in my quest for answers to these questions. In particular\, what exactly is the semantics of mCRL2 data types\, and what are the underlying choices an d their motivation (e.g. why do the data types have a model class semantics ?)
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jeroen-keiren-on-the-sema ntics-of-data-types-in-mcrl2/ END:VEVENT BEGIN:VEVENT UID:20220602T2243Z-1654209781.8821-EO-1151-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220531T100355Z LAST-MODIFIED:20220531T100355Z DTSTART;TZID=Europe/Amsterdam:20220519T124500 DTEND;TZID=Europe/Amsterdam:20220519T133000 SUMMARY: Jan Martens: Circular words\, Fibonacci words and it’s implication on partition refinement algorithms for bisimilarity DESCRIPTION: In this talk we will consider deterministic finite automata(DF As) with a singleton as alphabet. These rather restrictive machines have a strong connection with a very specific field of word combinatorics. In part icular we will show how the periodicity of the bouncing DVD logo[1](or a bi lliard ball) is related to these automata and generate so-called Fibonacci words. ...continue reading X-ALT-DESC;FMTTYPE=text/html:In this talk we will consider deterministi c finite automata(DFAs) with a singleton as alphabet. These rather restrict ive machines have a strong connection with a very specific field of word co mbinatorics. In particular we will show how the periodicity of the bouncing DVD logo[1](or a billiard ball) is related to these au tomata and generate so-called Fibonacci words. These words form an interest ing class of automata[2] that indicates that there might be more efficient methods to compute bisimilarity than the method of partition refinement.
[1] https://bouncingdvdlogo.com/
[2] "Hopcrof t's algorithm and cyclic automata" - Castiglione\, Restivo & Sciortino (20 08)
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-martens-circular-word s-fibonacci-words-and-its-implication-on-partition-refinement-algorithms-fo r-bisimilarity/ END:VEVENT BEGIN:VEVENT UID:20220602T2243Z-1654209781.9011-EO-1143-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220510T084643Z LAST-MODIFIED:20220510T084643Z DTSTART;TZID=Europe/Amsterdam:20220512T124500 DTEND;TZID=Europe/Amsterdam:20220512T133000 SUMMARY: Tom Franken: Cole’s Parallel Merge Sort DESCRIPTION: In this talk\, I shall explain Cole’s Parallel Merge Sorting A lgorithm\, which can sort lists of length n with O(n) processors in O(log n ) time. X-ALT-DESC;FMTTYPE=text/html:In this talk\, I shall explain Cole’s Para llel Merge Sorting Algorithm\, which can sort lists of length n with O(n) p rocessors in O(log n) time.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tom-franken-coles-paralle l-merge-sort/ END:VEVENT BEGIN:VEVENT UID:20220602T2243Z-1654209781.9187-EO-1138-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220426T154425Z LAST-MODIFIED:20220426T154957Z DTSTART;TZID=Europe/Amsterdam:20220429T150000 DTEND;TZID=Europe/Amsterdam:20220429T154500 SUMMARY: Dennis Rizvić: Making MCA easily understandable with mCRL2 DESCRIPTION: A model of multicopy semantics of low level memory operations is made in mCRL2. These memory operations can be relaxed\, release\, acquir e and sequentially consistent\, which determine whether these operations ca n be executed before or after surrounding instructions. Using this model it is investigated which semantics the read and write operations of Peterson’ s mutual exclusion algorithm must have to work ...continue reading X-ALT-DESC;FMTTYPE=text/html:A model of multicopy semantics of low leve l memory operations is made in mCRL2. These memory operations can be relaxe d\, release\, acquire and sequentially consistent\, which determine whether these operations can be executed before or after surrounding instructions. Using this model it is investigated which semantics the read and write ope rations of Peterson's mutual exclusion algorithm must have to work correctl y.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 12 GEO:51.447551;5.487453 ORGANIZER;CN="Jan Friso Groote":MAILTO:j.f.groote@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/dennis-rizvic-making-mca- easily-understandable-with-mcrl2/ END:VEVENT BEGIN:VEVENT UID:20220411T1333Z-1649683989.4943-EO-1133-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220411T131523Z LAST-MODIFIED:20220411T131523Z DTSTART;TZID=Europe/Amsterdam:20220414T124500 DTEND;TZID=Europe/Amsterdam:20220414T133000 SUMMARY: Olav Bunte: The impact of the channel layout in asynchronous commu nication DESCRIPTION: During our work on the asynchronous communication of OIL compo nents\, we started to wonder what the impact would be of changing the commu nication model. In this presentation I will present our current findings. I will shortly explain related work from Engels\, Mauw and Reniers on the hi erarchy of communication models that served as the basis ...continue readin g X-ALT-DESC;FMTTYPE=text/html:During our work on the asynchronous commun ication of OIL components\, we started to wonder what the impact would be o f changing the communication model. In this presentation I will present our current findings. I will shortly explain related work from Engels\, Mauw a nd Reniers on the hierarchy of communication models that served as the basi s and then show how we have adapted this to the context of FIFO systems. Di fferences between communication models is shown by an example. Lastly\, I s how in what cases confluence is preserved when switching to another communi cation model.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/olav-bunte-the-impact-of- the-channel-layout-in-asynchronous-communication/ END:VEVENT BEGIN:VEVENT UID:20220330T0800Z-1648627233.9838-EO-1130-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220330T075352Z LAST-MODIFIED:20220330T075352Z DTSTART;TZID=Europe/Amsterdam:20220331T124500 DTEND;TZID=Europe/Amsterdam:20220331T133000 SUMMARY: Anton Wijs: Inner-most Term Rewriting on GPUs DESCRIPTION: We present a way to implement term rewriting on a GPU. We do t his by letting the GPU repeatedly perform a massively parallel evaluation o f all subterms. We experimentally compared inner-most term rewriting with a relaxed form of inner-most rewriting\, and designed and experimented with two different garbage collection mechanisms\, to remove terms that are ...c ontinue reading X-ALT-DESC;FMTTYPE=text/html:We present a way to implement term rewriti ng on a GPU. We do this by letting the GPU repeatedly perform a massively p arallel evaluation of all subterms. We experimentally compared inner-most t erm rewriting with a relaxed form of inner-most rewriting\, and designed an d experimented with two different garbage collection mechanisms\, to remove terms that are no longer needed. We find that if the term rewrite systems exhibit sufficient internal parallelism\, GPU rewriting substantially outpe rforms the CPU. Both relaxed inner-most rewriting and garbage collection fu rther improves this performance. Since we expect that our implementation ca n be even further optimized\, and because in any case GPUs will become much more powerful in the future\, this suggests that GPUs are an interesting p latform for term rewriting. As term rewriting can be viewed as a universal programming language\, this also opens a route towards programming GPUs by term rewriting\, especially for irregular computations.
This is joint work with Johri van Eerd\, Jan Friso Groote\, Pieter Hijma\, Jan Martens a nd Muhammad Osama.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/anton-wijs-inner-most-ter m-rewriting-on-gpus/ END:VEVENT BEGIN:VEVENT UID:20220323T1149Z-1648036148.8061-EO-1129-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220323T090607Z LAST-MODIFIED:20220323T090607Z DTSTART;TZID=Europe/Amsterdam:20220324T124500 DTEND;TZID=Europe/Amsterdam:20220324T133000 SUMMARY: Michel Reniers: Partial-Order Reduction for Supervisory Controller Synthesis DESCRIPTION: A key challenge in the synthesis and subsequent analysis of su pervisory controllers is the impact of state-space explosion caused by conc urrency. The main bottleneck is often the memory needed to store the compos ition of plant and requirement automata and the resulting supervisor. Parti al-order reduction (POR) is a well-established technique that alleviates th is issue in the ...continue reading X-ALT-DESC;FMTTYPE=text/html:A key challenge in the synthesis and subse quent analysis of supervisory controllers is the impact of state-space expl osion caused by concurrency.
The main bottleneck is often the memory needed to store the composition of plant and requirement automata and the r esulting supervisor.
Partial-order reduction (POR) is a well-establis hed technique that alleviates this issue in the field of model checking. It does so by exploiting redundancy in the model with respect to the properti es of interest. For controller synthesis\, the functional properties of int erest are nonblockingness\, controllability\, and least-restrictiveness. Bu t also performance properties\, such as throughput and latency are of inter est. We propose POR on the input model that preserves both functional and p erformance properties in the synthesized supervisory controller. This impro ves scalability of the synthesis (and any subsequent performance analysis). Synthesis experiments show the effectiveness of the POR on a set of realis tic manufacturing system models.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/michel-reniers-partial-or der-reduction-for-supervisory-controller-synthesis/ END:VEVENT BEGIN:VEVENT UID:20220318T1029Z-1647599389.5454-EO-1128-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220318T094655Z LAST-MODIFIED:20220318T094655Z DTSTART;TZID=Europe/Amsterdam:20220317T124500 DTEND;TZID=Europe/Amsterdam:20220317T133000 SUMMARY: Rick Erkens: Term Rewriting based on Set Automaton Matching DESCRIPTION: In a previous talk I presented an efficient pattern matching a lgorithm based on the notion of set automaton. This matching algorithm can be exploited to implement efficient term rewriting procedures. These proced ures interleave pattern matching steps and rewriting steps\, and thus smoot hly integrate redex discovery and subterm replacement. In particular this m ethod is suitable to ...continue reading X-ALT-DESC;FMTTYPE=text/html:In a previous talk I presented an efficien t pattern matching algorithm based on the notion of set automaton. This mat ching algorithm can be exploited to implement efficient term rewriting proc edures. These procedures interleave pattern matching steps and rewriting st eps\, and thus smoothly integrate redex discovery and subterm replacement. In particular this method is suitable to implement the outermost rewriting strategy.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rick-erkens-term-rewritin g-based-on-set-automaton-matching/ END:VEVENT BEGIN:VEVENT UID:20220311T2218Z-1647037106.4088-EO-1120-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220311T140126Z LAST-MODIFIED:20220311T140126Z DTSTART;TZID=Europe/Amsterdam:20220303T124500 DTEND;TZID=Europe/Amsterdam:20220303T133000 SUMMARY: Flip van Spaendonck: Extending the Cones and Foci method DESCRIPTION: Proving two processes equivalent modulo branching bisimulation can be quite difficult and laborious. The cones and foci method seeks to s implify proving equivalence by assuming that\, in most implementations\, in ternal actions progress towards a state in which only externally visible ac tions are possible. In this talk\, we will discuss the original technique\, its incompleteness\, and ...continue reading X-ALT-DESC;FMTTYPE=text/html:Proving two processes equivalent modulo br anching bisimulation can be quite difficult and laborious.
The cones and foci method seeks to simplify proving equivalence by assuming that\, in most implementations\, internal actions progress towards a state in which only externally visible actions are possible.
In this talk\, we will discuss the original technique\, its incompleteness\, and extend upon it su ch that it can also be used to prove two processes equivalent modulo diverg ence-preserving branching bisimulation.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/flip-van-spaendonck-exten ding-the-cones-and-foci-method/ END:VEVENT BEGIN:VEVENT UID:20220228T1733Z-1646069608.4221-EO-1117-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220228T165025Z LAST-MODIFIED:20220228T165025Z DTSTART;TZID=Europe/Amsterdam:20220224T124500 DTEND;TZID=Europe/Amsterdam:20220224T133000 SUMMARY: Erik de Vink: On Spatial Logics and Spatial Bisimulation DESCRIPTION: Logics expressing spatial properties go back to Tarksi. Often their semantics is based on topological spaces. Cianza et al. propose to us e instead so-called closure spaces as underlying mathematical structure\, b ecause closure spaces comprise topological spaces as well as standard Kripk e frames. In this talk we exploit the fact that finite Kripke frames induce quasi-discrete ...continue reading X-ALT-DESC;FMTTYPE=text/html:Logics expressing spatial properties go ba ck to Tarksi. Often their semantics is based on topological spaces. Cianza et al. propose to use instead so-called closure spaces as underlying mathem atical structure\, because closure spaces comprise topological spaces as we ll as standard Kripke frames.
In this talk we exploit the fact that f inite Kripke frames induce quasi-discrete closure spaces\, strengthening th e connection of the closure operator and the accessibility relation. For th e corresponding Kripke models we relate a strong and a weak notion of bisim ulation to equivalence w.r.t. to a logic with a forward and a backward moda lity of “being near” and a forward and a backward path modality\, respectiv ely.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/erik-de-vink-on-spatial-l ogics-and-spatial-bisimulation/ END:VEVENT BEGIN:VEVENT UID:20220217T1426Z-1645107991.8986-EO-1101-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220217T140446Z LAST-MODIFIED:20220217T140446Z DTSTART;TZID=Europe/Amsterdam:20220217T124500 DTEND;TZID=Europe/Amsterdam:20220217T133000 SUMMARY: Bas Luttik: Axiomatising Sequencing and Signals DESCRIPTION: I will discuss a process algebra with constants for the deadlo cked and accepting processes\, action prefixing\, non-deterministic choice\ , sequencing\, signals and conditions. As Jos Baeten already mentioned in h is FSA colloquium talk of October 7\, 2021\, a process can be specified wit h a guarded recursive specification over this process algebra if\, and only if\, it ...continue reading X-ALT-DESC;FMTTYPE=text/html:I will discuss a process algebra with cons tants for the deadlocked and accepting processes\, action prefixing\, non-d eterministic choice\, sequencing\, signals and conditions. As Jos Baeten al ready mentioned in his FSA colloquium talk of October 7\, 2021\, a process can be specified with a guarded recursive specification over this process a lgebra if\, and only if\, it is stateless(ly?) bisimilar to the process ass ociated with a pushdown automaton. I’ll first briefly recap that correspond ence result\, and then proceed to present a sound and (ground-)complete axi omatisation of stateless bisimilarity in the context of the process algebra .
(The talk is based on ongoing joint work with Jos Baeten and Cesare Carissimo.)
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/bas-luttik-axiomatising-s equencing-and-signals/ END:VEVENT BEGIN:VEVENT UID:20220210T1207Z-1644494869.7364-EO-1097-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220210T111537Z LAST-MODIFIED:20220210T111537Z DTSTART;TZID=Europe/Amsterdam:20220210T124500 DTEND;TZID=Europe/Amsterdam:20220210T133000 SUMMARY: Mark Bouwman: Compositional state space generation DESCRIPTION: In the FormaSig project we translate communicating SysML state machines to mCRL2. Selecting and executing a transition consists of severa l transitions in the mCRL2 model. These internal state machines can be cons idered unobservable and are renamed to tau. Since many of these internal tr ansitions can be combined\, the bisimulation quotient is orders of magnitud e smaller ...continue reading X-ALT-DESC;FMTTYPE=text/html:In the FormaSig project we translat e communicating SysML state machines to mCRL2. Selecting and executing a tr ansition consists of several transitions in the mCRL2 model. These internal state machines can be considered unobservable and are renamed to tau. Sinc e many of these internal transitions can be combined\, the bisimulation quo tient is orders of magnitude smaller than the original LTS. However\, to be able to apply bisimulation reduction you first need to generate the entire LTS\, which is often not feasible for our models. In this talk I will pres ent a technique where we 1) compute the LTSs of individual state machines\, 2) reduce these LTSs modulo bisimulation and 3) combine the LTSs into the LTS representing the behaviour of the parallel composition of all state mac hines.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/mark-bouwman-compositiona l-state-space-generation/ END:VEVENT BEGIN:VEVENT UID:20220208T1005Z-1644314737.7867-EO-1096-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220208T090409Z LAST-MODIFIED:20220208T090409Z DTSTART;TZID=Europe/Amsterdam:20220203T124500 DTEND;TZID=Europe/Amsterdam:20220203T133000 SUMMARY: Jeroen Keiren: Towards effective unfolding of structured parameter s in mCRL2 DESCRIPTION: Data types such as lists and structured sorts enable the creat ion of concise process models in mCRL2. However\, static analysis tools suc h as constant elimination\, parameter elimination and sum elimination only consider process parameters as a single unit in their analysis. Therefore\, the added structure in process parameters negatively affects these static analysis techniques. Groote ...continue reading X-ALT-DESC;FMTTYPE=text/html:Data types such as lists and structured so rts enable the creation of concise process models in mCRL2. However\, stati c analysis tools such as constant elimination\, parameter elimination and s um elimination only consider process parameters as a single unit in their a nalysis. Therefore\, the added structure in process parameters negatively a ffects these static analysis techniques. Groote and Lisser introduced a tec hnique for flattening the structure of process parameters\, implemented in the mCRL2 tool lpsparunfold. This tools replaces a single structured proces s parameter by multiple parameters. In practice\, we often observe that\, a lthough parameters can be unfolded\, the static analysis tools do not benef it from this unfolding. In this talk I will explain parameter unfolding and show some examples of how it falls short in practice. I will also discuss ideas for improvements\, and the correctness criteria that parameter unfold ing should satisfy.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jeroen-keiren-towards-eff ective-unfolding-of-structured-parameters-in-mcrl2/ END:VEVENT BEGIN:VEVENT UID:20220127T1120Z-1643282446.3349-EO-1004-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220127T100833Z LAST-MODIFIED:20220127T100856Z DTSTART;TZID=Europe/Amsterdam:20220127T124500 DTEND;TZID=Europe/Amsterdam:20220127T133000 SUMMARY: Ferry Timmers: ASD to mCRL2\, translating and raking\, a retrospec tive DESCRIPTION: The past year I’ve been working on finalizing a translation fr om ASD (a specification language for control systems used by the industry) to mCRL2. This was not a trivial task since I was left without a formal spe cification of ASD. What ensued was reverse-engineering and lots of testing with purposefully built models. In the end\, ...continue reading X-ALT-DESC;FMTTYPE=text/html:The past year I've been working on finaliz ing a translation from ASD (a specification language for control systems us ed by the industry) to mCRL2. This was not a trivial task since I was left without a formal specification of ASD. What ensued was reverse-engineering and lots of testing with purposefully built models. In the end\, it require d a novel approach ('raking') to extract meaning from ASD generated source code\, which was quite successful. I will look back to this process\, and w hat we can draw from this.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/ferry-timmers-asd-to-mcrl 2-translating-and-raking-a-retrospective/ END:VEVENT BEGIN:VEVENT UID:20220120T1348Z-1642686480.7155-EO-999-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220120T134153Z LAST-MODIFIED:20220120T134153Z DTSTART;TZID=Europe/Amsterdam:20220120T124500 DTEND;TZID=Europe/Amsterdam:20220120T133000 SUMMARY: Tom Franken: Parallel Sorting Under Assumptions. DESCRIPTION: The talk will explore parallel sorting with the following assu mptions on the parallel processors: The processors can only save a constant amount of parameters. The processors need references to access other proce ssors. Notably\, I’ll take a look at sorting networks and the AKS network a nd at Richard Cole’s Parallel Merge Sort algorithm\, to see whether ...cont inue reading X-ALT-DESC;FMTTYPE=text/html:The talk will explore parallel sorting wit h the following assumptions on the parallel processors:
Notably\, I’ll ta ke a look at sorting networks and the AKS network and at Richard Cole’s Par allel Merge Sort algorithm\, to see whether they can be adapted to work wit h O(n) processors in O(log n) time under the two assumptions.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tom-franken-parallel-sort ing-under-assumptions/ END:VEVENT BEGIN:VEVENT UID:20220113T1649Z-1642092590.0289-EO-995-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220113T164944Z LAST-MODIFIED:20220113T164944Z DTSTART;TZID=Europe/Amsterdam:20220113T124500 DTEND;TZID=Europe/Amsterdam:20220113T133000 SUMMARY: Jan Martens: Lowerbounds for partition refinement algorithms that decide bisimilarity DESCRIPTION: Most of the algorithms that decide strong bisimilarity for LTS s can be classified as partition refinement algorithms. This includes the m ost efficient and well-known Paige-Tarjan algorithm. In recent work we esta blish an Omega((m+n) log n) lowerbound for the time complexity of these par tition refinement algorithms\, matching the time complexity of the Paige-Ta rjan algorithm. However there ...continue reading X-ALT-DESC;FMTTYPE=text/html:Most of the algorithms that decide strong bisimilarity for LTSs can be classified as partition refinement algorithms. This includes the most efficient and well-known Paige-Tarjan algorithm. In recent work we establish an Omega((m+n) log n) lowerbound for the time com plexity of these partition refinement algorithms\, matching the time comple xity of the Paige-Tarjan algorithm. However there is a catch: some techniqu es used for efficiently deciding bisimilarity on restricted LTSs that are d eterministic and only have a one-letter alphabet are not captured by the pa rtition refinement assumptions. Although it seems unlikely\, these techniqu es could improve the run-time complexity for deciding bisimilarity on less restricted LTSs. In this talk I'll discuss the lowerbound for partition ref inement algorithms and sketch the contradictory results that come with part ition refinement on restricted LTSs.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-martens-lowerbounds-f or-partition-refinement-algorithms-that-decide-bisimilarity/ END:VEVENT BEGIN:VEVENT UID:20220104T1455Z-1641308114.9492-EO-993-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220104T143935Z LAST-MODIFIED:20220104T143935Z DTSTART;TZID=Europe/Amsterdam:20220111T160000 DTEND;TZID=Europe/Amsterdam:20220111T173000 SUMMARY: Alexander Fedotov: Verification Techniques for xMAS DESCRIPTION: On January 11\, 2022\, Alexander Fedotov will defend his thesi s titled ‘Verification techniques for xMAS’\, which is available here. The defence will be streamed online via MS Teams. The link to the stream is ava ilable on request from a.fedotov@tue.nl. X-ALT-DESC;FMTTYPE=text/html:On January 11\, 2022\, Alexander Fedotov w ill defend his thesis titled ‘Verification techniques for xMAS’\, which is available here. The defence will be streamed online via MS Team s. The link to the stream is available on request from a.fedotov@tue.nl.
CATEGORIES:PhD Defence LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Jeroen Keiren":MAILTO:j.j.a.keiren@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/alexander-fedotov-verific ation-techniques-for-xmas/ END:VEVENT BEGIN:VEVENT UID:20220106T1446Z-1641480406.3088-EO-994-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20220106T143047Z LAST-MODIFIED:20220106T143047Z DTSTART;TZID=Europe/Amsterdam:20220106T124500 DTEND;TZID=Europe/Amsterdam:20220106T133000 SUMMARY: Tim Willemse: On quotients for equivalences on transition systems DESCRIPTION: There is a wealth of equivalence relations on labelled transit ion systems\; see\, e.g.\, Van Glabbeek’s linear-time branching-time spectr um. Some of these equivalences have found their way in tool sets such as mC RL2\, where they are used either to compare two transition systems\, or to reduce the size of a transition system. The latter is often ...continue rea ding X-ALT-DESC;FMTTYPE=text/html:There is a wealth of equivalence relations on labelled transition systems\; see\, e.g.\, Van Glabbeek's linear-time b ranching-time spectrum.
Some of these equivalences have found their w ay in tool sets such as mCRL2\, where they are used either to compare two t ransition systems\, or to reduce the size of a transition system. The latte r is often referred to as quotienting.
Somewhat surprisingly\, there does not appear to be a general recipe (yet) for obtaining a quotient for a given equivalence relation. To illustrate\, I'll explore the problem of qu otienting for several well-known equivalence relations.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tim-willemse-on-quotients -for-equivalences-on-transition-systems/ END:VEVENT BEGIN:VEVENT UID:20211213T1512Z-1639408335.6436-EO-987-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20211213T150740Z LAST-MODIFIED:20211213T150740Z DTSTART;TZID=Europe/Amsterdam:20211216T124500 DTEND;TZID=Europe/Amsterdam:20211216T133000 SUMMARY: Herman Geuvers: Programming with Higher Inductive Types DESCRIPTION: A relatively recent extension to type theory is “homotopy type theory”. This provides a new view on types\, where a type A is interpreted as a topological space\, a term t of type A is interpreted as a point in t he space and a proof of an equality\, p : q=t is interpreted as a ...contin ue reading X-ALT-DESC;FMTTYPE=text/html:A relatively recent extension to type theo ry is "homotopy type theory". This provides a new view on types\, where a t ype A is interpreted as a topological space\, a term t of type A is interpr eted as a point in the space and a proof of an equality\, p : q=t is interp reted as a path from q to t in the space A. This gives rise to new type the oretic principles (like "Univalence") and new type formers (like "Higher In ductive Types"). This allows to adequately formalize various concepts from algebraic topology\, but it is also relevant for computer science.
In the talk I will present some applications of homotopy type theory in compu ter science\, especially how to use Higher Inductive Types to write program s for "data types with laws"\, like the type of finite sets.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/herman-geuvers-programmin g-with-higher-inductive-types/ END:VEVENT BEGIN:VEVENT UID:20211207T1018Z-1638872307.0881-EO-986-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20211207T101524Z LAST-MODIFIED:20211207T101524Z DTSTART;TZID=Europe/Amsterdam:20211209T124500 DTEND;TZID=Europe/Amsterdam:20211209T133000 SUMMARY: Olav Bunte: Dynamic systems of communicating OIL components DESCRIPTION: In the previous presentation\, I showed how we can model async hronously communicating OIL components in mCRL2. Since then\, we have added a “new” operator in OIL which enables users to create new instances of com ponents dynamically. I show how we model the addition of this operator in m CRL2\, which required significant changes. Also\, I show ...continue readin g X-ALT-DESC;FMTTYPE=text/html:In the previous presentation\, I showed ho w we can model asynchronously communicating OIL components in mCRL2. Since then\, we have added a "new" operator in OIL which enables users to create new instances of components dynamically. I show how we model the addition o f this operator in mCRL2\, which required significant changes. Also\, I sho w how the notion of confluence can help with reducing the time to generate the state space in our context.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/olav-bunte-dynamic-system s-of-communicating-oil-components/ END:VEVENT BEGIN:VEVENT UID:20211203T1537Z-1638545852.6884-EO-983-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20211203T153212Z LAST-MODIFIED:20211203T153212Z DTSTART;TZID=Europe/Amsterdam:20211202T124500 DTEND;TZID=Europe/Amsterdam:20211202T133000 SUMMARY: Yousra Hafidi: Verifying Cordis models using mCRL2: some challenge s DESCRIPTION: Cordis models are industrial\, UML like models. There are some semantical differences compared to standard UML. We verify Cordis models u sing a translation to mCRL2. In order to verify such models\, we have to ov er several obstacles. In this talk\, I will discuss three main challenges t hat we face so far: (1) the accurate translation ...continue reading X-ALT-DESC;FMTTYPE=text/html:Cordis models are industrial\, UML like mo dels. There are some semantical differences compared to standard UML. We ve rify Cordis models using a translation to mCRL2. In order to verify such mo dels\, we have to over several obstacles. In this talk\, I will discuss thr ee main challenges that we face so far: (1) the accurate translation of Cor dis models into mCRL2\, (2) the specification of requirements about Cordis models using the mu-calculus\, and (3) scalability of the verification. I s how some examples of these challenges and discuss some solutions and result s. This work is part of the MACHINAIDE project (ITEA3\, No. 18030).
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/yousra-hafidi-verifying-c ordis-models-using-mcrl2-some-challenges/ END:VEVENT BEGIN:VEVENT UID:20211125T1104Z-1637838254.412-EO-981-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20211125T110302Z LAST-MODIFIED:20211125T110302Z DTSTART;TZID=Europe/Amsterdam:20211125T124500 DTEND;TZID=Europe/Amsterdam:20211125T133000 SUMMARY: Anton Wijs: Memory Efficient State Space Exploration on GPUs for C oncurrent State Machines with Data DESCRIPTION: GPUexplore is a tool that exploits the computational power of graphics processors to efficiently construct state spaces\, and on-the-fly check safety and liveness properties. Its current main practical limitation \, though\, is related to its input language. The tool accepts networks of Labelled Transition Systems\, which were initially useful to test whether s tate space could be ...continue reading X-ALT-DESC;FMTTYPE=text/html:GPUexplore is a tool that exploits the com putational power of graphics processors to efficiently construct state spac es\, and on-the-fly check safety and liveness properties. Its current main practical limitation\, though\, is related to its input language. The tool accepts networks of Labelled Transition Systems\, which were initially usef ul to test whether state space could be constructed efficiently on a GPU\, but when one wishes to encode data variables in such an input model\, the m odel quickly becomes unwieldy. In this talk\, I will present my recent resu lts on extending GPUexplore to support state machine models with data varia bles. To make this feasible\, a code generator has been created that given a model\, produces GPU code to interpret it. Furthermore\, to make the appr oach practical\, states are stored as binary trees to allow for sharing\, a nd Cleary table style compression is applied to more effectively use the GP U memory. The result is that state spaces consisting of billions of states can be explored in tens of seconds. This is the first work that investigate s state sharing and compression for GPUs.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/anton-wijs-memory-efficie nt-state-space-exploration-on-gpus-for-concurrent-state-machines-with-data/ END:VEVENT BEGIN:VEVENT UID:20211118T1839Z-1637260764.9039-EO-979-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20211118T182618Z LAST-MODIFIED:20211118T182700Z DTSTART;TZID=Europe/Amsterdam:20211119T140000 DTEND;TZID=Europe/Amsterdam:20211119T144500 SUMMARY: Koen Degeling: New algorithms and heuristics for solving Variabili ty Parity Games DESCRIPTION: Variability parity games are a recently proposed extension to well-known parity games that allow for verification of software product lin es (SPLs). We propose new algorithms for solving variability parity games b ased on the existing priority promotion and SCC decomposition\, and provide new heuristics for VPGs. We implemented these proposed algorithms\, as wel l as an algorithm ...continue reading X-ALT-DESC;FMTTYPE=text/html:Variability parity games are a recently pr oposed extension to well-known parity games that allow for verification of software product lines (SPLs). We propose new algorithms for solving variab ility parity games based on the existing priority promotion and SCC decompo sition\, and provide new heuristics for VPGs. We implemented these proposed algorithms\, as well as an algorithm based on Jurdzinski’s small progress measures. We compare existing algorithms for solving VPGs and the impact of different pre-processing steps and propose a method to generate random var iability parity games.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 2 GEO:51.447551;5.487453 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/koen-degeling-new-algorit hms-and-heuristics-for-solving-variability-parity-games/ END:VEVENT BEGIN:VEVENT UID:20211119T0932Z-1637314371.5765-EO-980-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20211119T092702Z LAST-MODIFIED:20211119T092702Z DTSTART;TZID=Europe/Amsterdam:20211118T124500 DTEND;TZID=Europe/Amsterdam:20211118T133000 SUMMARY: Jan Friso Groote: On the random structure of behavioural transitio n systems. DESCRIPTION: What is the structure of a transition system that represent th e behaviour of processes? We assumed that it was just an ordinary random gr aph\, but got odd results when predicting the sizes of state spaces generat ed by lps2lts. Viewing state spaces as parallel non-communicating random st ate spaces gave far better results. This also helps in ...continue reading X-ALT-DESC;FMTTYPE=text/html:What is the structure of a transition syst em that represent the behaviour of processes? We assumed that it was just a n ordinary random graph\, but got odd results when predicting the sizes of state spaces generated by lps2lts. Viewing state spaces as parallel non-com municating random state spaces gave far better results. This also helps in understanding why testing is so effective at times\, whereas some bugs inhe rently elude the testers.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-friso-groote-on-the-r andom-structure-of-behavioural-transition-systems/ END:VEVENT BEGIN:VEVENT UID:20211114T0932Z-1636882348.4082-EO-976-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20211114T093224Z LAST-MODIFIED:20211114T093224Z DTSTART;TZID=Europe/Amsterdam:20211111T124500 DTEND;TZID=Europe/Amsterdam:20211111T133000 SUMMARY: Anna Stramaglia: Semantics of UML State Machine Diagrams – overvie w and ambiguities DESCRIPTION: The Unified Modeling Language (UML)\, proposed by the Object M anagement Group (OMG)\, is a general purpose modeling language which became the standard for modeling system’ structure and behaviour. A UML model off ers different views of the system in the form of various diagrams. The talk ’s focus are UML State Machine Diagrams\, widely used to specify ...continu e reading X-ALT-DESC;FMTTYPE=text/html:The Unified Modeling Language (UML)\, prop osed by the Object Management Group (OMG)\, is a general purpose modeling l anguage which became the standard for modeling system' structure and behavi our. A UML model offers different views of the system in the form of variou s diagrams. The talk's focus are UML State Machine Diagrams\, widely used t o specify dynamic systems behaviours. The official semantics of UML State M achine Diagrams (and UML in general) is described in a semi-formal manner w here some bits are deliberately left undefined allowing different semantic interpretations. In this talk\, an overview of UML State Machine Diagrams i s given and some of the semantics ambiguities are pointed out\, describing how they have been faced and implemented in the literature.
CATEGORIES:Colloquium LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/anna-stramaglia-semantics -of-uml-state-machine-diagrams-overview-and-ambiguities/ END:VEVENT BEGIN:VEVENT UID:20211105T1012Z-1636107160.1083-EO-975-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20211105T101004Z LAST-MODIFIED:20211105T101004Z DTSTART;TZID=Europe/Amsterdam:20211104T124500 DTEND;TZID=Europe/Amsterdam:20211104T133000 SUMMARY: Thomas Neele: Compositional Learning of Synchronous Automata DESCRIPTION: The classical L* algorithm for learning DFAs runs in polynomia l time with the size of the DFA being learnt. However\, the DFA that repres ents a system consisting of multiple parallel components can grow exponenti ally in the number of components\; known as the state space explosion probl em. In this talk I will demonstrate how\, given the ...continue reading X-ALT-DESC;FMTTYPE=text/html:The classical L* algorithm for learning DF As runs in polynomial time with the size of the DFA being learnt. However\, the DFA that represents a system consisting of multiple parallel component s can grow exponentially in the number of components\; known as the state s pace explosion problem. In this talk I will demonstrate how\, given the abi lity to ask membership and equivalence queries for a (large) composite syst em\, we can learn its constituting components independently. This way\, we can avoid the state explosion and improve the performance of learning paral lel systems in practice.
CATEGORIES:Colloquium LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/thomas-neele-compositiona l-learning-of-synchronous-automata/ END:VEVENT BEGIN:VEVENT UID:20211025T1347Z-1635169675.3605-EO-974-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20211025T134020Z LAST-MODIFIED:20211025T134020Z DTSTART;TZID=Europe/Amsterdam:20211028T124500 DTEND;TZID=Europe/Amsterdam:20211028T133000 SUMMARY: Erik de Vink: In search of for stability: cancelativity for probab ilistic bisimulation DESCRIPTION: In an ongoing project on the complete axiomatization of branch ing probabilistic bisimulation\, we are currently focusing on a cancellatio n property. We see a route of proving the property by means of topological arguments which seems a bit far-fetched. As a possible alternative approach we propose the notion of a stable process. A process is stable ...continue reading X-ALT-DESC;FMTTYPE=text/html:In an ongoing project on the complete axio matization of branching probabilistic bisimulation\, we are currently focus ing on a cancellation property. We see a route of proving the property by m eans of topological arguments which seems a bit far-fetched. As a possible alternative approach we propose the notion of a stable process. A process i s stable if it doesn’t allow (partial) inert transitions. In the talk\, we discuss the setting\, the cancelativity\, property and whether each process will evolve into a stable processes eventually.
CATEGORIES:Colloquium LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/erik-de-vink-in-search-of -for-stability-cancelativity-for-probabilistic-bisimulation/ END:VEVENT BEGIN:VEVENT UID:20211019T1159Z-1634644749.4645-EO-973-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20211019T115026Z LAST-MODIFIED:20211019T115026Z DTSTART;TZID=Europe/Amsterdam:20211021T124500 DTEND;TZID=Europe/Amsterdam:20211021T133000 SUMMARY: Mark Bouwman: Verifying the EULYNX level crossing DESCRIPTION: Even though the behaviour of the EULYNX specification of the l evel crossing is conceptually quite simple\, the state space associated to the mCRL2 model is enormous. This Thursday I will present the reasons for t he state space explosion and the measures we took to drastically reduce it. Moreover\, I will present the verification efforts\, which ...continue rea ding X-ALT-DESC;FMTTYPE=text/html:Even though the behaviour of the EULYNX sp ecification of the level crossing is conceptually quite simple\, the state space associated to the mCRL2 model is enormous. This Thursday I will prese nt the reasons for the state space explosion and the measures we took to dr astically reduce it. Moreover\, I will present the verification efforts\, w hich includes some interesting results.
CATEGORIES:Colloquium LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/mark-bouwman-verifying-th e-eulynx-level-crossing/ END:VEVENT BEGIN:VEVENT UID:20211004T1425Z-1633357553.126-EO-961-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20211004T141656Z LAST-MODIFIED:20211004T141656Z DTSTART;TZID=Europe/Amsterdam:20211007T124500 DTEND;TZID=Europe/Amsterdam:20211007T133000 SUMMARY: Jos Baeten: Integration of automata theory and process theory DESCRIPTION: Since 2007\, I have been working on a project to integrate aut omata theory and process theory. The motivation is to present students with a model of a computer in a basic course on the foundations of computer sci ence\, that relates more closely to computers as they know them. A recent r esult in this project is ...continue reading X-ALT-DESC;FMTTYPE=text/html:Since 2007\, I have been working on a proj ect to integrate automata theory and process theory.
The motivation i s to present students with a model of a computer in a basic course on the f oundations of computer science\, that relates more closely to computers as they know them.
A recent result in this project is the addition of in teraction to the theorem of the equivalence of pushdown automata and contex t-free grammars.
CATEGORIES:Colloquium LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jos-baeten-integration-of -automata-theory-and-process-theory/ END:VEVENT BEGIN:VEVENT UID:20211004T1259Z-1633352351.6193-EO-958-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20211004T125550Z LAST-MODIFIED:20211004T125550Z DTSTART;TZID=Europe/Amsterdam:20211006T103000 DTEND;TZID=Europe/Amsterdam:20211006T111500 SUMMARY: Tom Buskens: Optimizing the code generator for OIL DESCRIPTION: OIL\, short for Open Interaction Language\, is a domain-specif ic language developed by Canon Production Printing B.V. It is a language th at can be used for specifying\, analyzing\, and implementing models of syst em behavior. The tooling created for OIL can generate C++ code from OIL spe cifications. Part of this generated code is a scheduler that schedules ...c ontinue reading X-ALT-DESC;FMTTYPE=text/html:OIL\, short for Open Interaction Language\ , is a domain-specific language developed by Canon Production Printing B.V. It is a language that can be used for specifying\, analyzing\, and impleme nting models of system behavior. The tooling created for OIL can generate C ++ code from OIL specifications. Part of this generated code is a scheduler that schedules so-called proactive events. The focus of this project is to improve this scheduler\; the aim is to reduce the number of computations n eeded to do the scheduling. We investigate basic scheduling improvement str ategies that do not need the collection of additional information. We also investigate scheduling strategies for which causal relations have to be gat hered from the OIL specifications. These scheduling strategies could be use d to skip the scheduling of events that are not needed. Another strategy th at we investigate alters the OIL specifications themselves to make them eas ier to schedule. In this report\, these strategies for improving the schedu ler are discussed\, verified\, and achieved results are listed.
CATEGORIES:MSc Defence LOCATION:Atlas 1.652 GEO:51.439265;5.478633 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tom-buskens-optimizing-th e-code-generator-for-oil/ END:VEVENT BEGIN:VEVENT UID:20210927T1545Z-1632757511.8999-EO-949-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210927T154425Z LAST-MODIFIED:20210927T154425Z DTSTART;TZID=Europe/Amsterdam:20210930T124500 DTEND;TZID=Europe/Amsterdam:20210930T133000 SUMMARY: Hans Zantema: The paint pot problem and common multiples in monoid s DESCRIPTION: On a finite sequence of paint pots the following steps are all owed: * Swap two consecutive non-empty pots. * If the two neighbours of a n on-empty pot are empty\, then divide the paint in the middle pot over the t wo neighbours\, after which these neighbours will be non-empty and the midd le one will be empty. ...continue reading X-ALT-DESC;FMTTYPE=text/html:On a finite sequence of paint pots the fol lowing steps are allowed:
* Swap two consecutive non-empty pots.
< p>* If the two neighbours of a non-empty pot are empty\, then divide the pa int in the middle pot over the two neighbours\, after which these neighbour s will be non-empty and the middle one will be empty. Also the reverse allo wed: if an empty pot has two neighbours of the same color\, the paint of th ese neigbours may be put in the middle pot.Is it possible to start b y a sequence in which the first four pots contain paint in four different c olors\, and get the first pot empty?
We will solve this remarkably ha rd problem by a remarkably simple solution\, and show its relation with mon oid theory\, and give generalizations of this problem\, partly solving an o pen problem presented by Jan Willem Klop.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/hans-zantema-the-paint-po t-problem-and-common-multiples-in-monoids/ END:VEVENT BEGIN:VEVENT UID:20210922T0914Z-1632302087.4147-EO-947-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210922T090727Z LAST-MODIFIED:20210922T090727Z DTSTART;TZID=Europe/Amsterdam:20210923T124500 DTEND;TZID=Europe/Amsterdam:20210923T133000 SUMMARY: Flip van Spaendonck: The Busy-Forbidden Protocol\, an efficient sh ared-exclusive access lock DESCRIPTION: Mutual exclusion algorithms such as Peterson’s\, make sure onl y a single thread can be present in the exclusive section at a given time. Similarly\, a shared-exclusive lock also provides a shared section\, in whi ch any number of threads can be present\, but only if no thread is present in the exclusive section. In this talk\, ...continue reading X-ALT-DESC;FMTTYPE=text/html:Mutual exclusion algorithms such as Peters
on's\, make sure only a single thread can be present in the exclusive secti
on at a given time. Similarly\, a shared-exclusive lock also provides a sha
red section\, in which any number of threads can be present\, but only if n
o thread is present in the exclusive section.
In this talk\, we will l
ook at one such shared-exclusive lock\, the Busy-Forbidden protocol\, the s
afety properties it satisfies\, and its performance.
Of special note i
s the low cost of entering and leaving the shared section using this protoc
ol. Making it especially useful for situations in which a significant amoun
t of the workload takes place in the shared section.
Tree automata are a generalisation of fini te automata over words. A tree automaton accepts a set of ranked ordered tr ees (terms if you will) just like a DFA accepts a set of words over an alph abet. Most interesting properties that regular languages enjoy\, carry over to regular tree languages in some sense. I will present some (dis)analogie s from the literature that I found interesting. In particular we will discu ss the conversion from nondeterministic bottom-up tree automata to determin istic ones\, and that there is no such conversion possible for top-down tre e automata. This negative result from the literature led to the research on pattern matching that I presented in my last talk.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rick-erkens-regular-tree- languages-and-disanalogies-with-regular-languages/ END:VEVENT BEGIN:VEVENT UID:20210912T1818Z-1631470736.6898-EO-944-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210912T144159Z LAST-MODIFIED:20210912T144159Z DTSTART;TZID=Europe/Amsterdam:20210913T140000 DTEND;TZID=Europe/Amsterdam:20210913T144500 SUMMARY: Anneke Huijsmans: Optimising parity game solvers using dynamic SCC maintenance DESCRIPTION: Two optimizations for Zielonka’s recursive algorithm for solvi ng parity games are investigated. The first optimization is partial re-deco mposition\, in which only the part of the graph containing vertices of SCCs which have 1 or more vertices removed will be re-decomposed. The second op timization is dynamic SCC maintenance\, which builds an SCC tree for each S CC ...continue reading X-ALT-DESC;FMTTYPE=text/html:Two optimizations for Zielonka's recursive algorithm for solving parity games are investigated. The first optimizatio n is partial re-decomposition\, in which only the part of the graph contain ing vertices of SCCs which have 1 or more vertices removed will be re-decom posed. The second optimization is dynamic SCC maintenance\, which builds an SCC tree for each SCC and then maintains those when vertices or edges are removed from the graph. An implementation in Java is made for 3 versions of Zielonka’s algorithm: the first version is Zielonka’s algorithm with Tarja n’s algorithm as described in literature. The second version is Zielonka’s algorithm with the use of partial re-decomposition. The third version is Zi elonka’s algorithm with dynamic SCC maintenance. The 3 versions are tested on various games. The conclusion from the tests is that Zielonka’s algorith m with partial re-decomposition gives the best improvement.
CATEGORIES:MSc Defence LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/anneke-huijsmans-optimisi ng-parity-game-solvers-using-dynamic-scc-maintenance/ END:VEVENT BEGIN:VEVENT UID:20210907T1705Z-1631034303.7152-EO-942-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210907T140353Z LAST-MODIFIED:20210907T140353Z DTSTART;TZID=Europe/Amsterdam:20210909T124500 DTEND;TZID=Europe/Amsterdam:20210909T133000 SUMMARY: Maurice Laveaux: On The Fly Solving for Symbolic Parity Games DESCRIPTION: In model checking we are interested in determining whether a p roperty is satisfied for a given system. In particular\, we want to know wh ether the property holds for the initial state of the system. We often see\ , especially while developing the model\, that properties are already viola ted in relatively few steps from the initial state. ...continue reading X-ALT-DESC;FMTTYPE=text/html:In model checking we are interested in det ermining whether a property is satisfied for a given system. In particular\ , we want to know whether the property holds for the initial state of the s ystem. We often see\, especially while developing the model\, that properti es are already violated in relatively few steps from the initial state. The refore\, we can attempt to verify the property while exploring the state sp ace of the system to find violations early. This approach has already succe ssfully been applied for explicit verification and we lift this to symbolic verification. We use symbolic parity games to encode the model checking qu estion. In this talk I will present the symbolic exploration technique exte nded with several partial solving methods.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/maurice-laveaux-on-the-fl y-solving-for-symbolic-parity-games/ END:VEVENT BEGIN:VEVENT UID:20210823T0705Z-1629702326.3747-EO-937-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210823T064919Z LAST-MODIFIED:20210823T064919Z DTSTART;TZID=Europe/Amsterdam:20210816T141500 DTEND;TZID=Europe/Amsterdam:20210816T150000 SUMMARY: Jasper Stam: Formal verification of an industrial PLC program in F unction Block Diagram and Structured Text DESCRIPTION: At Vitens\, the biggest drinking water company from the Nether lands\, most processes in extracting\, purifying and delivering drinking wa ter are automated using PLCs. In order to check PLC programs\, translation schemes for the programming languages Function Block Diagram and Structured Text are defined into an SMT solver. Using the SMT solver\, a set of typic al ...continue reading X-ALT-DESC;FMTTYPE=text/html:At Vitens\, the biggest drinking water com pany from the Netherlands\, most processes in extracting\, purifying and de livering drinking water are automated using PLCs. In order to check PLC pro grams\, translation schemes for the programming languages Function Block Di agram and Structured Text are defined into an SMT solver. Using the SMT sol ver\, a set of typical properties could be checked on a PLC program with th e size typical for PLC programs within Vitens. Most properties could be sho wn to hold (and all could be (dis)proven in matters of seconds)\, although virtually all initially formulated properties required fine tuning before t hey could be shown valid.
The overall conclusion is that the use of S MT solvers is very effective in analysing industrial scale PLC software.
CATEGORIES:MSc Defence LOCATION:Atlas 8.340 GEO:51.447488;5.486198 ORGANIZER;CN="Jan Friso Groote":MAILTO:j.f.groote@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jasper-stam-formal-verifi cation-of-an-industrial-plc-program-in-function-block-diagram-and-structure d-text/ END:VEVENT BEGIN:VEVENT UID:20210708T0941Z-1625737319.9669-EO-915-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210708T092728Z LAST-MODIFIED:20210708T092728Z DTSTART;TZID=Europe/Amsterdam:20210708T124500 DTEND;TZID=Europe/Amsterdam:20210708T133000 SUMMARY: Yousra Hafidi: On starvation freedom property of Peterson’s mutual exclusion algorithm for more than 2 processes DESCRIPTION: It is well known that Peterson’s algorithm for two processes i s starvation free. In this talk I study the generalization of Peterson’s al gorithm to N>2 processes using tournament trees. In particular\, I will sho w how\, contrary to the two processes version\, this algorithm is not starv ation free if we do not make any fairness assumptions. ...continue reading X-ALT-DESC;FMTTYPE=text/html:It is well known that Peterson's algorithm for two processes is starvation free. In this talk I study the generalizat ion of Peterson's algorithm to N>2 processes using tournament trees. In par ticular\, I will show how\, contrary to the two processes version\, this al gorithm is not starvation free if we do not make any fairness assumptions. To overcome this issue\, I propose a new variation of the algorithm that en forces starvation freedom and bounded overtaking without imposing any fairn ess assumption.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Anna Stramaglia":MAILTO:a.stramaglia@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/yousra-hafidi-on-starvati on-freedom-property-of-petersons-mutual-exclusion-algorithm-for-more-than-2 -processes/ END:VEVENT BEGIN:VEVENT UID:20210630T1736Z-1625074614.304-EO-911-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210630T173004Z LAST-MODIFIED:20210630T173004Z DTSTART;TZID=Europe/Amsterdam:20210701T124500 DTEND;TZID=Europe/Amsterdam:20210701T133000 SUMMARY: Herman Geuvers: “Computer Assisted Mathematical Proofs: using the computer to verify computers” DESCRIPTION: A “Proof Assistant” is a computer program that allows users to create complete mathematical proofs\, interactively with the computer\, wh ere the computer checks each small reasoning step. In this way one obtains the utmost guarantee of correctness. I will outline how Proof Assistants wo rk\, how they are used to verify mathematical proofs and computer systems. Verifying a proof with a ...continue reading X-ALT-DESC;FMTTYPE=text/html:A "Proof Assistant" is a computer program that allows users to create complete mathematical proofs\, interactively wi th the computer\, where the computer checks each small reasoning step. In t his way one obtains the utmost guarantee of correctness.
I will outli ne how Proof Assistants work\, how they are used to verify mathematical pro ofs and computer systems. Verifying a proof with a PA is a lot of work\, bu t as mathematical proofs get more and more difficult and complex there is a n increasing use of PAs for mathematical proofs. Also for critical computer components it will pay off to verify them completely using a PA.
We will discuss why one would trust a PA and their limitations in use\, which basically rest on the limitations of proof automation. It has recently beco me clear that Machine Learning provides methods that apply very well to spe eding up proof automation.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/herman-geuvers-computer-a ssisted-mathematical-proofs-using-the-computer-to-verify-computers/ END:VEVENT BEGIN:VEVENT UID:20210616T1216Z-1623845770.0863-EO-903-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210616T120307Z LAST-MODIFIED:20210616T120307Z DTSTART;TZID=Europe/Amsterdam:20210617T124500 DTEND;TZID=Europe/Amsterdam:20210617T133000 SUMMARY: Olav Bunte: Asynchronously communicating OIL components DESCRIPTION: In this presentation we focus on how we envision OIL component s to communicate with each other. We define the behaviour of communicating OIL components using a formalism found in literature\, namely the FIFO syst em\, and we show how this can be modelled in mCRL2. Also\, we show some typ ical unwanted behaviour in asynchronous communication and ...continue readi ng X-ALT-DESC;FMTTYPE=text/html:In this presentation we focus on how we en vision OIL components to communicate with each other. We define the behavio ur of communicating OIL components using a formalism found in literature\, namely the FIFO system\, and we show how this can be modelled in mCRL2. Als o\, we show some typical unwanted behaviour in asynchronous communication a nd discuss the difficulty of checking whether such behaviour is present.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/olav-bunte-asynchronously -communicating-oil-components/ END:VEVENT BEGIN:VEVENT UID:20210601T1707Z-1622567262.5411-EO-900-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210601T165700Z LAST-MODIFIED:20210601T165700Z DTSTART;TZID=Europe/Amsterdam:20210603T124500 DTEND;TZID=Europe/Amsterdam:20210603T133000 SUMMARY: Jeroen Keiren: Understanding Mutual Exclusion Algorithms using mCR L2’s counterexamples DESCRIPTION: Last week\, Myrthe Spronck discussed mutual exclusion algorith ms when using safe registers. In this week’s talk\, I will stick with the t opic of mutual exclusion algorithms\, but switch back to using atomic regis ters. I will introduce Dekker’s algorithm for mutual exclusion. According t o Dijkstra\, this is the first algorithm to solve mutual exclusion between two ...continue reading X-ALT-DESC;FMTTYPE=text/html:Last week\, Myrthe Spronck discussed mutua l exclusion algorithms when using safe registers. In this week’s talk\, I w ill stick with the topic of mutual exclusion algorithms\, but switch back t o using atomic registers. I will introduce Dekker’s algorithm for mutual ex clusion. According to Dijkstra\, this is the first algorithm to solve mutua l exclusion between two processes.
In my talk\, I will first discuss some of the historical context of Dekker’s algorithm. Subsequently\, follow ing Dijkstra’s 1962 paper “Over de sequentialiteit van procesbeschrijvingen ” (EWD35)\, I will show how some simpler attempts fail\, and\, guided by th e counterexamples that mCRL2 produces\, develop this into Dekker’s algorith m.
During the talk\, I will address several desirable requirements of mutual exclusion algorithms\, as well as some modelling considerations.
This talk is based on the mCRL2 tutorial presentation “Designin g Distributed Algorithms in mCRL2” that Jan Friso Groote and I will present at DisCoTec on June 14.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jeroen-keiren-understandi ng-mutual-exclusion-algorithms-using-mcrl2s-counterexamples/ END:VEVENT BEGIN:VEVENT UID:20210526T1110Z-1622027409.8389-EO-899-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210526T110432Z LAST-MODIFIED:20210526T110432Z DTSTART;TZID=Europe/Amsterdam:20210527T124500 DTEND;TZID=Europe/Amsterdam:20210527T133000 SUMMARY: Myrthe Spronck: Safe registers and Aravind’s BLRU algorithm in mCR L2 DESCRIPTION: For my bachelor research project\, supervised by Bas Luttik\, I set out to verify Aravind’s bounded least recently used (BLRU) algorithm for mutual exclusion using mCRL2. An interesting property of Aravind’s algo rithm is that it can ensure mutual exclusion even when the registers used a re safe\, rather than atomic. In order to verify this property\, ...continu e reading X-ALT-DESC;FMTTYPE=text/html:For my bachelor research project\, supervi sed by Bas Luttik\, I set out to verify Aravind's bounded least recently us ed (BLRU) algorithm for mutual exclusion using mCRL2. An interesting proper ty of Aravind's algorithm is that it can ensure mutual exclusion even when the registers used are safe\, rather than atomic. In order to verify this p roperty\, I modelled the behavior of safe registers in mCRL2. I will presen t this model\, and show the effects of safe registers by demonstrating how Peterson's algorithm violates mutual exclusion when safe registers are used . I will also discuss the results of verifying Aravind's algorithm.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/myrthe-spronck-safe-regis ters-and-aravinds-blru-algorithm-in-mcrl2/ END:VEVENT BEGIN:VEVENT UID:20210518T1720Z-1621358454.2723-EO-898-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210518T171551Z LAST-MODIFIED:20210518T171551Z DTSTART;TZID=Europe/Amsterdam:20210520T124500 DTEND;TZID=Europe/Amsterdam:20210520T133000 SUMMARY: Mark Bouwman: A formalisation of SysML state machines in mCRL2 DESCRIPTION: My talk this Thursday will consist of three parts. (10 -15 min ) I will practice my talk for FORTE\, where I will present the paper “A for malisation of SysML state machines in mCRL2”. This paper reports on a forma lisation of the semi-formal modelling language SysML in the formal language mCRL2. The formalisation focuses on a ...continue reading X-ALT-DESC;FMTTYPE=text/html:My talk this Thursday will consist of thre e parts.
In this talk\, I will discuss McMillan's a lgorithm for fully SAT-based unbounded symbolic model checking. The method is based on computing Craig interpolants. In terms of performance\, the alg orithm is substantially more efficient than BDD-based model checking. I wil l also explain how we modify McMillan's algorithm to analyze the backward r eachability of initial states from final states in a transition system.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/alexander-fedotov-mcmilla ns-algorithm-for-sat-based-unbounded-model-checking/ END:VEVENT BEGIN:VEVENT UID:20210428T0944Z-1619603045.5809-EO-893-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210428T094236Z LAST-MODIFIED:20210428T094236Z DTSTART;TZID=Europe/Amsterdam:20210429T124500 DTEND;TZID=Europe/Amsterdam:20210429T133000 SUMMARY: Flip van Spaendonck: Understanding std::memory_order in C++11 usin g MCA semantics DESCRIPTION: C++11 introduced many tools to write safe multi threaded code. One of those tools are the std::memory_orders\, which specify how memory a ccesses\, including regular\, non-atomic memory accesses\, are to be ordere d around atomic operations. Understanding these memory orders can be quite a complex situation\, specifically when different memory orders are combine d. We will try to make ...continue reading X-ALT-DESC;FMTTYPE=text/html:C++11 introduced many tools to write safe multi threaded code. One of those tools are the std::memory_orders\, which specify how memory accesses\, including regular\, non-atomic memory accesse s\, are to be ordered around atomic operations. Understanding these memory orders can be quite a complex situation\, specifically when different memor y orders are combined. We will try to make sense of these by using mCRL2 to analyze all traces possible under MCA semantics. A set of semantics in whi ch all threads perceive memory accesses in the same order.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/flip-van-spaendonck-under standing-stdmemory_order-in-c11-using-mca-semantics/ END:VEVENT BEGIN:VEVENT UID:20210421T0838Z-1618994304.8504-EO-890-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210421T081344Z LAST-MODIFIED:20210421T081344Z DTSTART;TZID=Europe/Amsterdam:20210422T150000 DTEND;TZID=Europe/Amsterdam:20210422T154500 SUMMARY: Geert van Ieperen: Visualisation of large Labelled Transition Syst ems DESCRIPTION: A formal model describes the behaviour of a program\, protocol or other system. The properties of this model can be verified\, such that we can prove these properties with absolute certainty. Visualising the stat e space of a formal model is an important tool for their development. This thesis focuses on visualising such a state space ...continue reading X-ALT-DESC;FMTTYPE=text/html:A formal model describes the behaviour of
a program\, protocol or other system. The properties of this model can be v
erified\, such that we can prove these properties with absolute certainty.
Visualising the state space of a formal model is an important tool for thei
r development. This thesis focuses on visualising such a state space as a g
raph.
We present a layout technique that provides efficient graph layo
uts for graphs with over 10000 nodes\, discuss the effectiveness of differe
nt edge shapes\, and present multiple ways of highlighting the structure of
the graph. We present a rendering procedure capable of rendering graphs la
rger than 10000 nodes while supporting these techniques. We implemented the
se techniques in a tool and use it to measure the performance of the techni
ques. The effectiveness of these techniques is analysed using a survey on t
he tool in a few practical applications.
We conclude that these techn iques can effectively be used in one visualisation to assist in formal mode l development.
The tool is registered on DOI 10.5281/zenodo.4680646 p> CATEGORIES:MSc Defence LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/geert-van-ieperen-visuali sation-of-large-labelled-transition-systems/ END:VEVENT BEGIN:VEVENT UID:20210421T0626Z-1618986404.2484-EO-887-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210421T055419Z LAST-MODIFIED:20210421T055419Z DTSTART;TZID=Europe/Amsterdam:20210422T124500 DTEND;TZID=Europe/Amsterdam:20210422T133000 SUMMARY: Anna Stramaglia: Deadlock in packet switching networks DESCRIPTION: A deadlock in a packet switching network is a state in which o ne or more messages have not yet reached their target\, yet cannot progress any further. We formalize three different notions of deadlock in the conte xt of packet switching networks\, to which we refer as global\, local and w eak deadlock. We establish the precise ...continue reading X-ALT-DESC;FMTTYPE=text/html:
A deadlock in a packet switching network i s a state in which one or more messages have not yet reached their target\, yet cannot progress any further. We formalize three different notions of d eadlock in the context of packet switching networks\, to which we refer as global\, local and weak deadlock. We establish the precise relations betwee n these notions\, and prove they characterize different sets of deadlocks. Moreover\, we implement checking of deadlock freedom of packet switching ne tworks using the symbolic model checker nuXmv. We show experimentally that the implementation is effective at finding subtle deadlock situations in pa cket switching networks.
This is joint work with Jeroen J.A. Keiren a nd Hans Zantema
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/anna-stramaglia-deadlock- in-packet-switching-networks-2/ END:VEVENT BEGIN:VEVENT UID:20210414T1814Z-1618424068.1355-EO-883-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210414T172217Z LAST-MODIFIED:20210414T172217Z DTSTART;TZID=Europe/Amsterdam:20210415T124500 DTEND;TZID=Europe/Amsterdam:20210415T133000 SUMMARY: Muhammad Osama: SAT Solving with GPU Accelerated Inprocessing DESCRIPTION: Since 2013\, the leading SAT solvers in the SAT competition al l use inprocessing\, which unlike preprocessing\, interleaves search with s implifications. However\, applying inprocessing frequently can still be a b ottle neck\, i.e.\, for hard or large formulas. In this work\, we introduce the first attempt to parallelize inprocessing on GPU architectures. As mem ory is a scarce ...continue reading X-ALT-DESC;FMTTYPE=text/html:Since 2013\, the leading SAT solvers in th e SAT competition all use inprocessing\, which unlike preprocessing\, inter leaves search with simplifications. However\, applying inprocessing frequen tly can still be a bottle neck\, i.e.\, for hard or large formulas. In this work\, we introduce the first attempt to parallelize inprocessing on GPU a rchitectures. As memory is a scarce resource in GPUs\, we present new space -efficient data structures and devise a data-parallel garbage collector. It runs in parallel on the GPU to reduce memory consumption and improves memo ry access locality. Our new parallel variable elimination algorithm is twic e as fast as previous work. In experiments our new solver ParaFROST solves many benchmarks faster on the GPU than its sequential counterparts.
T his is a joint work with Armin Biere (Johannes Kepler University).
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/muhammad-osama-sat-solvin g-with-gpu-accelerated-inprocessing/ END:VEVENT BEGIN:VEVENT UID:20210406T1604Z-1617725083.7118-EO-882-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210406T160114Z LAST-MODIFIED:20210406T160114Z DTSTART;TZID=Europe/Amsterdam:20210408T124500 DTEND;TZID=Europe/Amsterdam:20210408T133000 SUMMARY: Rick Erkens: A Set Automaton to Locate All Pattern Matches in a Te rm DESCRIPTION: Term pattern matching is the problem of finding all pattern ma tches in a subject term\, given a set of patterns. It is a fundamental algo rithmic problem in for instance automated theorem proving and term rewritin g. We present a set automaton solution for the term pattern matching proble m that is based on match set derivatives where ...continue reading X-ALT-DESC;FMTTYPE=text/html:Term pattern matching is the problem of fi nding all pattern matches in a subject term\, given a set of patterns. It i s a fundamental algorithmic problem in for instance automated theorem provi ng and term rewriting. We present a set automaton solution for the term pat tern matching problem that is based on match set derivatives where each fun ction symbol in the subject pattern is visited exactly once. The algorithm allows for various traversal patterns over the subject term and is particul arly suited to search the subject term in parallel using a large number of simultaneously running threads.
This is joint work with Jan Friso Gro ote.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rick-erkens-a-set-automat on-to-locate-all-pattern-matches-in-a-term/ END:VEVENT BEGIN:VEVENT UID:20210401T1030Z-1617273054.9847-EO-868-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210401T100503Z LAST-MODIFIED:20210401T100503Z DTSTART;TZID=Europe/Amsterdam:20210401T124500 DTEND;TZID=Europe/Amsterdam:20210401T133000 SUMMARY: Bas Luttik: Equationally axiomatising parallel composition DESCRIPTION: I will overview some results pertaining to the (equational) ax iomatisation of interleaving parallel composition. X-ALT-DESC;FMTTYPE=text/html:I will overview some results pertaining to the (equational) axiomatisation of interleaving parallel composition.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/bas-luttik-equationally-a xiomatising-parallel-composition/ END:VEVENT BEGIN:VEVENT UID:20210323T1401Z-1616508077.175-EO-866-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210323T135413Z LAST-MODIFIED:20210323T135413Z DTSTART;TZID=Europe/Amsterdam:20210325T124500 DTEND;TZID=Europe/Amsterdam:20210325T133000 SUMMARY: Jan Friso Groote: Partition refinement algorithms for strong bisim ulation are Omega(n log n) DESCRIPTION: A question haunting me for a while is whether the O(m log n) a lgorithm for strong bisimulation is optimal. We found a family of graphs th at shows that any reasonable partition refinement algorithm is necessarily Omega(n log n)\, n being the number of states\, steps to calculate strong bisimulation. This appeared to answer the question. ...continue reading X-ALT-DESC;FMTTYPE=text/html:A question haunting me for a while is whet her the O(m log n) algorithm for strong bisimulation is optimal. We found a family of graphs that shows that any reasonable partition refinement algo rithm is necessarily Omega(n log n)\, n being the number of states\, steps to calculate strong bisimulation. This appeared to answer the question. Bu t some results from the literature -- digged up by Jan Martens -- puts our finding in a strange light. There is a family of graphs for which branchin g bisimulation can be calculated linearly\, but which still is Omega(n log n) when partition refinement is applied. These results need to be understoo d further\, but in extremo this suggests that partition refinement is not s uch a good idea after all.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-friso-groote-partitio n-refinement-algorithms-for-strong-bisimulation-are-omegan-log-n/ END:VEVENT BEGIN:VEVENT UID:20210316T1334Z-1615901655.9545-EO-865-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210316T131937Z LAST-MODIFIED:20210316T131937Z DTSTART;TZID=Europe/Amsterdam:20210318T124500 DTEND;TZID=Europe/Amsterdam:20210318T133000 SUMMARY: Hans Zantema: Complexity of Simon’s problem in classical sense DESCRIPTION: Simon’s problem is a standard example of a problem that is exp onential in classical sense\, while it admits a linear solution in quantum computing. It is about a function f for which it is given that a unique non -zero vector s exists for which f(x) = f(x xor s) for all x. The goal is to find s. The ...continue reading X-ALT-DESC;FMTTYPE=text/html:Simon's problem is a standard example of a problem that is exponential in classical sense\, while it admits a linear solution in quantum computing. It is about a function f for which it is giv en that a unique non-zero vector s exists for which f(x) = f(x xor s) for a ll x. The goal is to find s. The exponential lower bound for the classical sense assumes that f only admits black box access. In this presentation we investigate classical complexity when f is given by a standard representati on like a circuit. We focus on finding the vector space of all vectors s fo r which f(x) = f(x xor s) for all x\, for any given f. Two main results are : (1) if f is given by any circuit\, then checking whether this vector spac e contains a non-zero element is NP-hard\, and (2) when restricting to BDDs \, then a basis of this vector space can be computed in polynomial time.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/hans-zantema-complexity-o f-simons-problem-in-classical-sense/ END:VEVENT BEGIN:VEVENT UID:20210309T2106Z-1615324017.97-EO-862-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210309T203356Z LAST-MODIFIED:20210309T203356Z DTSTART;TZID=Europe/Amsterdam:20210311T124500 DTEND;TZID=Europe/Amsterdam:20210311T133000 SUMMARY: Jan Martens: Solving bisimilarity and the relational coarsest part ition problem in sub-linear parallel time seems hard DESCRIPTION: It is known that deciding bisimilarity is a P-complete problem . This means it is thought of as a problem that is inherently sequential an d hard to solve in parallel. Despite this fact efforts have been made to co nstruct increasingly efficient parallel algorithms. In a previous colloquiu m I presented a parallel algorithm that decides bisimilarity in ...continue reading X-ALT-DESC;FMTTYPE=text/html:It is known that deciding bisimilarity is a P-complete problem. This means it is thought of as a problem that is inhe rently sequential and hard to solve in parallel. Despite this fact efforts have been made to construct increasingly efficient parallel algorithms. In a previous colloquium I presented a parallel algorithm that decides bisimil arity in O(n) time on n+m parallel processors that are allowed to concurren tly read and write. In this talk I will discuss what the parallel intractab ility of P-complete problems means\, and what this means for further improv ements of the algorithm.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-martens-solving-bisim ilarity-and-the-relational-coarsest-partition-problem-in-sub-linear-paralle l-time-seems-hard/ END:VEVENT BEGIN:VEVENT UID:20210302T0924Z-1614677043.9817-EO-858-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210302T090624Z LAST-MODIFIED:20210302T090624Z DTSTART;TZID=Europe/Amsterdam:20210304T124500 DTEND;TZID=Europe/Amsterdam:20210304T133000 SUMMARY: Erik de Vink: On Quantum Process Algebra DESCRIPTION: After a quick intro to quantum computing addressing Deutsch’s problem\, we turn to quantum teleportation and look into what may be needed to handle such with a tool like (probabilistic) mCRL2. X-ALT-DESC;FMTTYPE=text/html:After a quick intro to quantum computing a ddressing Deutsch's problem\, we turn to quantum teleportation and look int o what may be needed to handle such with a tool like (probabilistic) mCRL2.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/erik-de-vink-on-quantum-p rocess-algebra/ END:VEVENT BEGIN:VEVENT UID:20210222T1459Z-1614005974.4316-EO-857-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210222T140535Z LAST-MODIFIED:20210222T140535Z DTSTART;TZID=Europe/Amsterdam:20210225T124500 DTEND;TZID=Europe/Amsterdam:20210225T133000 SUMMARY: Maurice Laveaux: Antichain Based Algorithm for Fair Testing DESCRIPTION: The notion of refinement plays an important role in software e ngineering. It is the basis of a stepwise development methodology in which the correctness of a system can be established by proving\, or computing\, that a system refines its specification. There are many refinement relation s described in the literature. Fair testing is the coarsest liveness-preser ving ...continue reading X-ALT-DESC;FMTTYPE=text/html:The notion of refinement plays an importan t role in software engineering. It is the basis of a stepwise development m ethodology in which the correctness of a system can be established by provi ng\, or computing\, that a system refines its specification. There are many refinement relations described in the literature. Fair testing is the coar sest liveness-preserving refinement relation that is a precongruence for a CSP inspired process algebra. The main feature of fair testing is that it a bstracts from divergences in the same way as Milner's observation congruenc e\, and as a result is also strictly coarser than observation congruence. T his can be seen as a built-in fairness assumption. Fair testing has been sh own to be decidable\, but its algorithm is not yet practical. In this prese ntation I will describe fair testing and the algorithm. Furthermore\, I wil l also explain the applicability of antichains to improve the efficiency of the algorithm.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/maurice-laveaux-antichain -based-algorithm-for-fair-testing/ END:VEVENT BEGIN:VEVENT UID:20210208T0837Z-1612773463.0867-EO-856-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210208T080235Z LAST-MODIFIED:20210208T080235Z DTSTART;TZID=Europe/Amsterdam:20210211T124500 DTEND;TZID=Europe/Amsterdam:20210211T133000 SUMMARY: Ferry Timmers: State-space exploration of generated system control lers DESCRIPTION: Model-driven system engineering is a practice also employed in the design of controllers for cyber-physical systems. The method allows co ntrollers to be modelled and verified before they are implemented in softwa re\, allowing potential glitches and design flaws to be uncovered\, before they emerge in the time and resource intensive testing phase. Some of the m odelling ...continue reading X-ALT-DESC;FMTTYPE=text/html:Model-driven system engineering is a pract ice also employed in the design of controllers for cyber-physical systems. The method allows controllers to be modelled and verified before they are i mplemented in software\, allowing potential glitches and design flaws to be uncovered\, before they emerge in the time and resource intensive testing phase. Some of the modelling formalisms used by the industry allow controll ers to be automatically generated from models\, which aides in the preventi on of human error in the implementation of such systems. The question arise s whether the behavior of the generated software components is equivalent t o those of the specified models\, as the semantics of such systems might no t always be trivially deduced. The topic of this talk is about how to solve this problem. It will present a way to deduct the state-space from impleme nted software controllers\, and the intricacies of this approach. It will g ive the context of where this question arose\, and what its contribution ca n potentially be for the industry and in a general setting. It will conclud e with a few examples to shows the approach in working order\, and if you a re still reading this\, given there is enough time and interest in the subj ect a more interactive demonstration.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/ferry-timmers-state-space -exploration-of-generated-system-controllers/ END:VEVENT BEGIN:VEVENT UID:20210201T1201Z-1612180908.5806-EO-855-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210201T113516Z LAST-MODIFIED:20210201T113516Z DTSTART;TZID=Europe/Amsterdam:20210204T124500 DTEND;TZID=Europe/Amsterdam:20210204T133000 SUMMARY: Yousra Hafidi: Modeling and Improved Verification of Reconfigurabl e Discrete Event Systems using R-TNCESs Formalism DESCRIPTION: Reconfigurability is a concept that appeared recently in sever al areas including manufacturing\, aerospace\, medical\, robotic\, and tele communication systems. This concept provides systems with an aspect of flex ibility allowing them to easily adapt with their external environment durin g their working process. Reconfiguration provides many advantages to variou s existing systems. However\, by adopting such aspect\, some issues ...cont inue reading X-ALT-DESC;FMTTYPE=text/html:Reconfigurability is a concept that appear ed recently in several areas including manufacturing\, aerospace\, medical\ , robotic\, and telecommunication systems. This concept provides systems wi th an aspect of flexibility allowing them to easily adapt with their extern al environment during their working process.
Reconfiguration provides many advantages to various existing systems. However\, by adopting such as pect\, some issues can be confronted and deserve to be resolved. In this ta lk we will treat problems related to the modeling and the verification of r econfigurable systems using reconfigurable timed net condition/event system s (R-TNCES) formalism. R-TNCESs formalism is a modular extension of Petri n ets enriched with conditions\, events\, time\, and reconfiguration constrai nts.
This work was part of my PhD research works in collaboration wit h University of Tunis El Manar\, Jinan University\, and University of Moham ed Khider Biskra.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/yousra-hafidi-modeling-an d-improved-verification-of-reconfigurable-discrete-event-systems-using-r-tn cess-formalism/ END:VEVENT BEGIN:VEVENT UID:20210127T0905Z-1611738336.9966-EO-854-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210127T084714Z LAST-MODIFIED:20210128T074133Z DTSTART;TZID=Europe/Amsterdam:20210128T124500 DTEND;TZID=Europe/Amsterdam:20210128T133000 SUMMARY: Jeroen Keiren: Peterson’s mutual exclusion algorithm for n process es DESCRIPTION: When talking about mutual exclusion\, many textbooks start by introducing Peterson’s algorithm for two processes. The algorithm looks ver y simple\, but upon closer inspection its behaviour is deceptively subtle. Less commonly known are extensions of Peterson’s algorithm to n processes. In this talk\, I will look at such generalisations. I will also describe ho w these ...continue reading X-ALT-DESC;FMTTYPE=text/html:When talking about mutual exclusion\, many textbooks start by introducing Peterson’s algorithm for two processes. The algorithm looks very simple\, but upon closer inspection its behaviour is deceptively subtle. Less commonly known are extensions of Peterson’s algori thm to n processes. In this talk\, I will look at such generalisations. I w ill also describe how these processes can be modelled and verified using mC RL2.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jeroen-keiren-petersons-m utual-exclusion-algorithm-for-n-processes/ END:VEVENT BEGIN:VEVENT UID:20210120T1007Z-1611137276.958-EO-853-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210120T083413Z LAST-MODIFIED:20220203T111044Z DTSTART;TZID=Europe/Amsterdam:20210121T124500 DTEND;TZID=Europe/Amsterdam:20210121T133000 SUMMARY: Mark Bouwman: Direct formalization of EULYNX SysML in mCRL2 DESCRIPTION: FormaSig aims to strengthen railway signalling standardization processes with the use of formal methods. The concrete approach investigat ed in FormaSig is to derive formal mCRL2 models from existing SysML specifi cations. These formal models are then used for two distinct purposes: (i) c hecking whether the original standard satisfies the requirements that are i mposed upon them\, and ...continue reading X-ALT-DESC;FMTTYPE=text/html:F ormaSig aims to strengthen railway signalling standardization processes with the use of formal methods. The concrete approach investigated in Form aSig is to derive formal mCRL2 models from existing SysML specifications. T hese formal models are then used for two distinct purposes: (i) checking wh ether the original standard satisfies the requirements that are imposed upo n them\, and (ii) performing automated testing of implementations. In this talk I will present our approach to developing a translation from SysML to mCRL2. We define the semantics of state machines directly in mCRL2. The mod el containing the generic semantics then only needs to be completed with a specific instantiation of state machines by encoding the structure of these state machines in the mCRL2 data language. We have a prototype translation tool that takes a number of SysML diagrams and produces such a configurati on.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/mark-bouwman-direct-forma lization-of-eulynx-sysml-in-mcrl2/ END:VEVENT BEGIN:VEVENT UID:20210104T1500Z-1609772418.1938-EO-845-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20210104T145819Z LAST-MODIFIED:20210104T145819Z DTSTART;TZID=Europe/Amsterdam:20210107T124500 DTEND;TZID=Europe/Amsterdam:20210107T133000 SUMMARY: Wieger Wesselink: Symbolic Reachability using LDDs DESCRIPTION: The mCRL2 toolset contains several applications in which compu ting the set of reachable states of a transition relation plays a role. For example in state space generation and in solving a PBES. List decision dia grams (LDDs) can be used to store sets of states and transitions in a compa ct manner. This has been demonstrated by ...continue reading X-ALT-DESC;FMTTYPE=text/html:The mCRL2 toolset contains several applica tions in which computing the set of reachable states of a transition relati on plays a role. For example in state space generation and in solving a PBE S. List decision diagrams (LDDs) can be used to store sets of states and tr ansitions in a compact manner. This has been demonstrated by Jaco van de Po l and Stefan Blom in 2008\, and these ideas have been implemented in the LT SMin toolset. We are currently applying these ideas in the mCRL2 toolset as well\, using the LDD functionality of the Sylvan library made by Tom van D ijk. In this talk I will explain the main ideas behind it\, and describe a reachability algorithm that uses LDDs.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/wieger-wesselink-symbolic -reachability-using-ldds/ END:VEVENT BEGIN:VEVENT UID:20201214T2337Z-1607989074.2596-EO-843-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20201214T210526Z LAST-MODIFIED:20201214T213157Z DTSTART;TZID=Europe/Amsterdam:20201217T124500 DTEND;TZID=Europe/Amsterdam:20201217T133000 SUMMARY: Herman Geuvers: Relating Apartness and Bisimulation DESCRIPTION: We have studied the dual of bisimulation: the notion of “apart ness”. Intuitively\, two elements are apart if there is a positive way to d istinguish them. Apartness is the dual of bisimilarity in a precise categor ical sense: apartness is an initial algebra and gives rise to an induction principle. In the talk we will focus on the inductive nature of ...continue reading X-ALT-DESC;FMTTYPE=text/html:We take a look at k-DFAOs\, which are Dete
rministic Finite Automata with Output with a special property: each k-DFAO
represents a k-automatic sequence a\, an infinite sequence in which the i-t
h element is the output the automata produces for the k-ary representation
of i. Given any k-automatic sequence a\, we define their complexity ||a||k
as the size of the smallest possible k-DFAO representing our sequence and s
imilarly the reverse complexity ||a||Rk for the right to left representatio
n of k-ary numbers.
To be more specific\, we look at local changes f t
o our sequences\, that only change a finite amount of elements\, and find a
n upper bound for the complexities ||f(a)||k and ||f(a)||Rk\, when applied
to an arbitrary sequence a.
We then use SAT/SMT solvers to prove that
these upper bounds can not be further improved\, thus establishing a lower
bound as well.
We also create an algorithm for minimizing any k-DFAO\,
which will give us a more efficient way of getting ||a||k than using a SAT
/SMT solver.
Whenever two Labelled Transition Systems ( LTSs) are behaviourally inequivalent to each other\, one may be interested why this is the case. Using a modal logic that characterises such a behavio ural equivalence one can create formulae that distinguish these two LTSs\, exposing the reason for the inequivalence. In this presentation I will desc ribe the work by Rance Cleaveland on computing distinguishing formulae for strong bisimulation and the work by Henri Korver on computing distinguishin g formulae for branching bisimulation. Also\, I will compare the work of Ko rver with an alternate logic for characterising branching bisimulation and propose ideas to (possibly) improve on Korvers work.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/olav-bunte-evidence-for-b ehavioural-inequivalence/ END:VEVENT BEGIN:VEVENT UID:20201123T1055Z-1606128910.1675-EO-837-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20201123T104809Z LAST-MODIFIED:20201123T104809Z DTSTART;TZID=Europe/Amsterdam:20201127T130000 DTEND;TZID=Europe/Amsterdam:20201127T134500 SUMMARY: Wouter Schols: Verification of an iterative implementation of Tarj an’s algorithm for Strongly Connected Components using Dafny DESCRIPTION: Tarjan’s algorithm for strongly connected components is used i n the mCRL2 toolset. This algorithm finds all strongly connected components or SCC in a directed graph. Here an SCC is a maximal set of nodes such tha t there exists a directed path between all nodes in the set. The normal imp lementation of Tarjan’s algorithm uses recursion. ...continue reading X-ALT-DESC;FMTTYPE=text/html:Tarjan’s algorithm for strongly connected components is used in the mCRL2 toolset. This algorithm finds all strongly connected components or SCC in a directed graph. Here an SCC is a maximal s et of nodes such that there exists a directed path between all nodes in the set. The normal implementation of Tarjan’s algorithm uses recursion. Unfor tunately\, this implementation can cause stack overflow problems if the gra phs are sufficiently large. The mCRL2 toolset uses an iterative implementat ion of Tarjan’s algorithm to circumvent this problem. In this project both the recursive and the much more complex iterative algorithm have been prove n correct using the verification language Dafny. For the recursive algorith m a previous proof was already available. However this proof was nearing th e limits of Dafny’s capabilities\, which caused unstable results after mino r changes. As part of this work techniques have been introduced to guarante e much more stable proofs.
CATEGORIES:MSc Defence LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Wieger Wesselink":MAILTO:J.W.Wesselink@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/wouter-schols-verificatio n-of-an-iterative-implementation-of-tarjans-algorithm-for-strongly-connecte d-components-using-dafny/ END:VEVENT BEGIN:VEVENT UID:20201123T2049Z-1606164582.8019-EO-838-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20201123T182733Z LAST-MODIFIED:20201123T182821Z DTSTART;TZID=Europe/Amsterdam:20201126T124500 DTEND;TZID=Europe/Amsterdam:20201126T133000 SUMMARY: Anton Wijs: “Lock and Fence When Needed: State Space Exploration + Static Analysis = Improved Fence and Lock Insertion” DESCRIPTION: When targeting modern parallel hardware architectures\, constr ucting correct and high-performing software is complex and time-consuming. In particular\, reorderings of memory accesses that violate intended sequen tially consistent behaviour are a major source of bugs. Applying synchronis ation mechanisms to repair these should be done sparingly\, as they negativ ely impact performance. In the past\, both static analysis approaches and t echniques ...continue reading X-ALT-DESC;FMTTYPE=text/html:When targeting modern parallel hardware ar chitectures\, constructing correct and high-performing software is complex and time-consuming. In particular\, reorderings of memory accesses that vio late intended sequentially consistent behaviour are a major source of bugs. Applying synchronisation mechanisms to repair these should be done sparing ly\, as they negatively impact performance.
In the past\, both static
analysis approaches and techniques based on explicit-state model checking
have been proposed to identify where synchronisation fences have to be plac
ed in a program. The former are fast\, but the latter more precise\, as the
y tend to insert fewer fences. Unfortunately\, the model checking technique
s suffer a form of state space explosion that is even worse than the
t
raditional one.
In this work\, we propose a technique using a combina tion of state space exploration and static analysis. This combination is in terms of precision comparable to purely model checking-based techniques\, but it reduces the state space explosion problem to the one typically seen in model checking. Furthermore\, experiments show that the combination freq uently outperforms both purely model checking and static analysis technique s. In addition\, we have added the capability to check for atomicity violat ions\, which is another major source of bugs.
I have talked about thi s topic before in the FSA colloquium\, but since then\, this work has been accepted\, and presented last week\, at iFM 2020. Since the previous talk\, a new implementation of the approach has been constructed\, and new experi ments have been performed\, providing interesting insights.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/anton-wijs-lock-and-fence -when-needed-state-space-exploration-static-analysis-improved-fence-and-loc k-insertion/ END:VEVENT BEGIN:VEVENT UID:20201118T0828Z-1605688091.6695-EO-820-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20201118T082605Z LAST-MODIFIED:20201118T082605Z DTSTART;TZID=Europe/Amsterdam:20201119T124500 DTEND;TZID=Europe/Amsterdam:20201119T133000 SUMMARY: Tim Willemse: On Recursive Algorithms for Solving Parity Games DESCRIPTION: Parity Games are infinite duration\, two-player graph games. S uch games play an important role in verification\, satisfiability and synth esis. In recent years\, several quasi-polynomial time algorithms for solvin g parity games have appeared. One of the more recent ones\, by Pawel Parys\ , is based on the classical divide-and-conquer algorithm by McNaughton/Ziel onka. In this talk\, I will ...continue reading X-ALT-DESC;FMTTYPE=text/html:Parity Games are infinite duration\, two-p layer graph games. Such games play an important role in verification\, sati sfiability and synthesis. In recent years\, several quasi-polynomial time a lgorithms for solving parity games have appeared. One of the more recent on es\, by Pawel Parys\, is based on the classical divide-and-conquer algorith m by McNaughton/Zielonka. In this talk\, I will reiterate this classical al gorithm and shed some light on the modifications that allow for achieving t he quasi-polynomial runtime. While the big leap in runtime complexity is im pressive\, experiments indicate that the performance of this new algorithm sucks. This seems to be characteristic for all 'improved' algorithms.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tim-willemse-on-recursive -algorithms-for-solving-parity-games/ END:VEVENT BEGIN:VEVENT UID:20201109T1542Z-1604936552.185-EO-818-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20201109T133936Z LAST-MODIFIED:20201109T133936Z DTSTART;TZID=Europe/Amsterdam:20201112T124500 DTEND;TZID=Europe/Amsterdam:20201112T133000 SUMMARY: Rick Erkens: “Bloom’s Cool Congruence Formats for Weak Behavioral Equivalences: Make Branching Bisimilarity a Congruence Again” DESCRIPTION: The field of process algebra provides a way to model (concurre nt) processes algebraically. The syntax of a process calculus is described by an algebraic signature and the semantics is described by a set of operat ional rules. Strong bisimilarity and branching bisimilarity are two well kn own behavioral equivalences on processes. Informally two processes are stro ngly bisimilar ...continue reading X-ALT-DESC;FMTTYPE=text/html:The field of process algebra provides a wa y to model (concurrent) processes algebraically. The syntax of a process ca lculus is described by an algebraic signature and the semantics is describe d by a set of operational rules. Strong bisimilarity and branching bisimila rity are two well known behavioral equivalences on processes. Informally tw o processes are strongly bisimilar if they cannot be distinguished by an ex ternal observer. Branching bisimilarity is a weaker equivalence that abstra cts from internal transitions.
In many process calculi\, strong bisim ilarity is a congruence. Informally this enables one to reason algebraicall y about strong bisimilarity of two processes. The GSOS format is a family o f process calculi for which strong bisimilarity is guaranteed to be a congr uence. Unfortunately branching bisimilarity is not a congruence for a GSOS language in general. To mitigate this issue\, Bard Bloom introduced languag e formats by putting extra restrictions on the GSOS rules. In particular th e BB-cool format guarantees that branching bisimilarity is a congruence.
In this talk I will demonstrate why congruence is a complicated issue i n weak behavioral equivalences. We will discuss the GSOS format\, the BB-co ol format\, and a sketch of a proof by Rob van Glabbeek that shows why the restrictions on GSOS are enough to make branching bisimilarity a congruence again. Currently I am studying this topic at a more abstract level with Ba s Luttik and Jurriaan Rot\, which I will briefly comment on.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rick-erkens-blooms-cool-c ongruence-formats-for-weak-behavioral-equivalences-make-branching-bisimilar ity-a-congruence-again/ END:VEVENT BEGIN:VEVENT UID:20201103T1323Z-1604409790.0398-EO-817-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20201103T121939Z LAST-MODIFIED:20201103T121939Z DTSTART;TZID=Europe/Amsterdam:20201105T124500 DTEND;TZID=Europe/Amsterdam:20201105T133000 SUMMARY: Maurice Laveaux: Decompositional Minimisation of Monolithic Proces ses DESCRIPTION: Compositional minimisation can be an effective technique to re duce the state space explosion problem. This technique considers a parallel composition of several processes. In its simplest form\, each sequential p rocess is replaced by an abstraction\, simpler than the corresponding proce ss while still preserving the property that is checked. However\, this tech nique cannot be applied in ...continue reading X-ALT-DESC;FMTTYPE=text/html:Compositional minimisation can be an effec tive technique to reduce the state space explosion problem. This technique considers a parallel composition of several processes. In its simplest form \, each sequential process is replaced by an abstraction\, simpler than the corresponding process while still preserving the property that is checked. However\, this technique cannot be applied in a setting where parallel com position is first translated to a non-deterministic sequential monolithic p rocess. The advantage of this monolithic process is that it facilitates sta tic analysis of global behaviour. We present a technique that considers a m onolithic process with data and decomposes it into two processes where each process defines behaviour for a subset of the parameters of the monolithic process. In this talk I will also highlight some of the practical consider ations of this technique and the results that have been achieved so-far.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/maurice-laveaux-decomposi tional-minimisation-of-monolithic-processes/ END:VEVENT BEGIN:VEVENT UID:20201026T1634Z-1603730050.347-EO-815-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20201026T150407Z LAST-MODIFIED:20201026T150407Z DTSTART;TZID=Europe/Amsterdam:20201029T124500 DTEND;TZID=Europe/Amsterdam:20201029T133000 SUMMARY: Bas Luttik: Off-the-shelf automated analysis of liveness propertie s for just paths DESCRIPTION: The motivation for the research that I will discuss in my talk is a claim by van Glabbeek and Höfner that CCS-like process algebras are n ot powerful enough to establish correctness of Peterson’s algorithm\, and\, in particular\, prove the required liveness property. The culprit\, accord ing to van Glabbeek and Höfner\, is that for these process ...continue read ing X-ALT-DESC;FMTTYPE=text/html:Bram Hooimeijer did do his research for th e master thesis at ASML. The question was whether it is possible to constru ct models for the behaviour of software out of a single trace using informa tion about the software architecture. Bram showed that it was possible to c ome up with a reasonable model that could even be finetuned by analysing th e obtained model\, for instance for deadlocks and by adding meta informatio n to the model for instance about the allowed ordering of certain events in the software.
CATEGORIES:MSc Defence LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Jan Friso Groote":MAILTO:j.f.groote@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/bram-hooimeijer-model-inf erence-for-legacy-software-in-component-based-architectures/ END:VEVENT BEGIN:VEVENT UID:20201012T1431Z-1602513063.2887-EO-804-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20201012T141507Z LAST-MODIFIED:20201012T141507Z DTSTART;TZID=Europe/Amsterdam:20201015T124500 DTEND;TZID=Europe/Amsterdam:20201015T133000 SUMMARY: Hans Zantema: Complexity of puzzles DESCRIPTION: For several puzzles (often from Japanese origin) the following can be investigated How to solve them\, either by backtracking or by SAT/S MT? How to generate them? How to prove NP-completeness? Currently I supervi se several students in Nijmegen doing projects in this area. In this talk t hese projects will be discussed. X-ALT-DESC;FMTTYPE=text/html:For several puzzles (often from Japanese o rigin) the following can be investigated
Currently I supervise several stude nts in Nijmegen doing projects in this area. In this talk these projects wi ll be discussed.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/hans-zantema-complexity-o f-puzzles/ END:VEVENT BEGIN:VEVENT UID:20201006T1718Z-1602004688.1285-EO-801-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20201006T170020Z LAST-MODIFIED:20201006T170020Z DTSTART;TZID=Europe/Amsterdam:20201008T124500 DTEND;TZID=Europe/Amsterdam:20201008T133000 SUMMARY: Jan Martens: A parallel algorithm for bisimulation partition refin ement DESCRIPTION: The notion of strong bisimilarity on LTSs is an equivalence th at expresses whether states have the same behavior. In their pioneering wor k Kanellakis and Smolka proposed a partition refinement algorithm that find s the coarsest partition of bisimilar states. We propose a parallel version of this algorithm that performs this partition refinement on an LTS with . ..continue reading X-ALT-DESC;FMTTYPE=text/html:The notion of strong bisimilarity on LTSs is an equivalence that expresses whether states have the same behavior. In their pioneering work Kanellakis and Smolka proposed a partition refinement algorithm that finds the coarsest partition of bisimilar states. We propos e a parallel version of this algorithm that performs this partition refinem ent on an LTS with n states and m labels in O(n) steps on max(n\,m) CRCW PR AM processors. Additionally\, it does not perform more work than the sequen tial counterpart.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-martens-a-parallel-al gorithm-for-bisimulation-partition-refinement/ END:VEVENT BEGIN:VEVENT UID:20200922T1259Z-1600779549.1628-EO-783-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200922T120007Z LAST-MODIFIED:20200922T120007Z DTSTART;TZID=Europe/Amsterdam:20200924T124500 DTEND;TZID=Europe/Amsterdam:20200924T133000 SUMMARY: Anna Stramaglia: Deadlock in packet switching networks DESCRIPTION: Deadlock is a historically well-known bug pattern in computer systems where\, in the most general sense\, a system reaches a state in whi ch it cannot progress any further. In this work\, we investigate deadlocks in packet switching networks with a deterministic routing function. We form alize three different notions of deadlock\, namely\, global\, local and wea k ...continue reading X-ALT-DESC;FMTTYPE=text/html:Deadlock is a historically well-known bug pattern in computer systems where\, in the most general sense\, a system re aches a state in which it cannot progress any further. In this work\, we in vestigate deadlocks in packet switching networks with a deterministic routi ng function. We formalize three different notions of deadlock\, namely\, gl obal\, local and weak deadlock. We formally define the relation between the se different notions. Moreover\, we formalize packet switching networks in nuXmv and verify them for each of the notions of deadlock\, based on the th eory proposed. We also explore the translation of these networks to xMAS. p> CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/anna-stramaglia-deadlock- in-packet-switching-networks/ END:VEVENT BEGIN:VEVENT UID:20200921T0703Z-1600671807.3597-EO-780-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200921T065519Z LAST-MODIFIED:20200921T065638Z DTSTART;TZID=Europe/Amsterdam:20200921T133000 DTEND;TZID=Europe/Amsterdam:20200921T140000 SUMMARY: Nikita Golovliov: Verification of Multiprocessor System Memory Mod el DESCRIPTION: In a shared memory multiprocessor system\, the memory model de termines the outcome of read operations at any time\, given a partial order of memory operations induced by processor-issued writes and reads. A memor y model may pertain to high-level language semantics or hardware program ex ecution. This thesis focuses on verification of hardware memory model confo rmance\, i.e. ...continue reading X-ALT-DESC;FMTTYPE=text/html:
In a shared memory multiprocessor system \, the memory model determines the outcome of read operations at any time\, given a partial order of memory operations induced by processor-issued wri tes and reads. A memory model may pertain to high-level language semantics or hardware program execution.
This thesis focuses on verification of hardware memory model conformance\, i.e. given a hardware specification an d memory model\, verifying whether all the possible executions of any progr am on the hardware are permitted by the memory model\, and vice versa.
< p>View-based definitions of a memory model require existence of a certain o rder between memory modification and observation events\, called view. If\, and only if\, there exists a view that is consistent with the memory model for every execution of any program by a multiprocessor system\, then the m ultiprocessor system conforms to the memory model.Given a generic hi gh-level processor-memory interface\, we provide a translation of the relea se consistency hardware memory model to a set of linear temporal logic prop erties. We analyse the correctness of the translation and the deficiencies of the approach. We discuss the relation of the hardware memory model to th e memory model of C++.
We provide an example multiprocessor system de sign that is used for verifica- tion of memory model conformance\, and the guidelines for applying the properties to a real-world system.
We present a general method to verify an m CRL2 model of a multiprocessor with respect to memory consistency and prove the correctness of this method.
Consequently\, we present a way to r eformulate most memory models that are defined in terms of serial views usi ng observations. We prove that any execution of a program by a multiprocess or satisfying this formulation is memory consistent under the original memo ry consistency model. We then express this as a set of predicates on the tr aces allowed by an mCRL2 model and prove that if all traces allowed by the mCRL2 model satisfy these predicates\, then all executions allowed by the mCRL2 model are consistent with respect to the original memory consistency model.
We also formulate a subset of the C++ memory model in terms of serial views. We prove that if such that serial view exists for every exec ution allowed by a multiprocessor\, then every execution allowed by that pr ocessor is consistent under the C++ memory model. The predicates on mCRL2 t races are then expressed as mu-calculus formulas and are used to verify an mCRL2 model of a multiprocessor with respect to local consistency and cache consistency\, which are subsets of the serial view representing the C++ me mory model. We use these mu-calculus formulas to benchmark the verification an example CPU model using the mCRL2 toolset.
CATEGORIES:MSc Defence LOCATION:Atlas 2.215 GEO:52.374540;4.897976 ORGANIZER;CN="Jeroen Keiren":MAILTO:j.j.a.keiren@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/wessel-sinnema-verifying- memory-consistency-in-mcrl2/ END:VEVENT BEGIN:VEVENT UID:20200921T0641Z-1600670472.1065-EO-778-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200921T063251Z LAST-MODIFIED:20200921T063251Z DTSTART;TZID=Europe/Amsterdam:20200917T124500 DTEND;TZID=Europe/Amsterdam:20200917T133000 SUMMARY: Alexander Fedotov: Effective System Level Liveness Verification DESCRIPTION: The language xMAS has been designed by Intel with the purpose of modelling and verification of hardware. Recently\, the language was exte nded with finite state machines to make it more expressive. Furthermore\, i t was shown how to prove liveness of such extended xMAS networks. Unfortuna tely\, we demonstrate that the proof technique is unsound. We provide ...co ntinue reading X-ALT-DESC;FMTTYPE=text/html:The language xMAS has been designed by Int el with the purpose of modelling and verification of hardware. Recently\, t he language was extended with finite state machines to make it more express ive. Furthermore\, it was shown how to prove liveness of such extended xMAS networks. Unfortunately\, we demonstrate that the proof technique is unsou nd. We provide an alternative approach which we have carefully proven to be correct. Moreover\, we show that our approach scales very well\, which mak es it possible to prove liveness properties at the system level. In particu lar\, we show that using our approach\, it is possible to verify a power co ntrol architecture composed of 1299 state machines representing 50 power do mains where each domain contains 5 master and 5 slave devices. Proving live ness of this system takes less than 10 minutes.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/alexander-fedotov-effecti ve-system-level-liveness-verification/ END:VEVENT BEGIN:VEVENT UID:20200826T1606Z-1598457975.1003-EO-680-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200826T135859Z LAST-MODIFIED:20200826T140530Z DTSTART;TZID=Europe/Amsterdam:20200916T160000 DTEND;TZID=Europe/Amsterdam:20200916T173000 SUMMARY: Thomas Neele: Reductions for Parity Games and Model Checking DESCRIPTION: On September 16th 2020\, Thomas Neele will defend his thesis t itle ‘Reductions for Parity Games and Model Checking’\, which is available here. The defence will be streamed online via MS Teams. The link to the str eam is available on request from t.s.neele@tue.nl. X-ALT-DESC;FMTTYPE=text/html:On September 16th 2020\, Thomas Neele will defend his thesis title 'Reductions for Parity Games and Model Checking'\, which is available here. The defence will be str eamed online via MS Teams. The link to the stream is available on request f rom t.s.neele@tue.nl.
CATEGORIES:PhD Defence LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Thomas Neele":MAILTO:t.s.neele@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/phd-defense-thomas-neele/ END:VEVENT BEGIN:VEVENT UID:20200909T1004Z-1599645857.195-EO-774-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200909T081944Z LAST-MODIFIED:20200909T081944Z DTSTART;TZID=Europe/Amsterdam:20200910T124500 DTEND;TZID=Europe/Amsterdam:20200910T133000 SUMMARY: Ferry Timmers: ASD\, the good\, the bad\, and the ugly. DESCRIPTION: The ASD framework\, a proprietary standard for model driven en gineering\, has been around for over a decade now. It provides a toolset to design and verify control systems\, and has successfully been employed by the industry. We look at the formalisms to see what makes it tick\; the adv antages of using ASD\, its limitations\, and ...continue reading X-ALT-DESC;FMTTYPE=text/html:The ASD framework\, a proprietary standard for model driven engineering\, has been around for over a decade now. It p rovides a toolset to design and verify control systems\, and has successful ly been employed by the industry. We look at the formalisms to see what mak es it tick\; the advantages of using ASD\, its limitations\, and problems t hat may rise. The goal is to give a more general impression of ASD\, rather than a detailed explanation.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/ferry-timmers-asd-the-goo d-the-bad-and-the-ugly/ END:VEVENT BEGIN:VEVENT UID:20200817T1242Z-1597668139.8341-EO-766-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200817T122247Z LAST-MODIFIED:20200817T122709Z DTSTART;TZID=Europe/Amsterdam:20200818T153000 DTEND;TZID=Europe/Amsterdam:20200818T160000 SUMMARY: Sebastiaan Verhoek: SMT solver verification of ladder logic in a p roduction environment (Tata Steel) DESCRIPTION: It is shown how to verify requirements on the PLC code in use at Tata Steel. A translator from PLC programs to the input language for SMT solvers has been written. Subsequently\, requirements on some of the large st PLC programs available at Tata steel have been written down\, and their validity on the software has ...continue reading X-ALT-DESC;FMTTYPE=text/html:It is shown how to verify requirements on the PLC code in use at Tata Steel. A translator from PLC programs to the in put language for SMT solvers has been written. Subsequently\, requirements on some of the largest PLC programs available at Tata steel have been writt en down\, and their validity on the software has been validated.
CATEGORIES:MSc Defence LOCATION:Atlas 2.320 GEO:52.374540;4.897976 ORGANIZER;CN="Jan Friso Groote":MAILTO:j.f.groote@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/sebastiaan-verhoek-smt-so lver-verification-of-ladder-logic-in-a-production-environment-tata-steel/ END:VEVENT BEGIN:VEVENT UID:20200624T0925Z-1592990707.2201-EO-736-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200624T083345Z LAST-MODIFIED:20200624T083345Z DTSTART;TZID=Europe/Amsterdam:20200625T124500 DTEND;TZID=Europe/Amsterdam:20200625T133000 SUMMARY: Anton Wijs: Term Rewriting on GPUs DESCRIPTION: We are interested in finding a programming model to employ the power of Graphics Processing Units (GPUs) more easily than via traditional programming. We propose to express computations in the form of term rewrit e systems. We present a way to implement term rewriting on a GPU. We do thi s by letting the GPU repeatedly perform ...continue reading X-ALT-DESC;FMTTYPE=text/html:This presentation is the third in the cont ext of our work on OIL\, a language for modelling control software\, where we have enabled the use of model checking via a translation to mCRL2. In th is presentation we will dive a bit deeper into the complexity of this trans lation. Also\, to test the feasibility of our methods we have applied them to a second model from Canon\, which is considerably larger than the one di scussed in the previous presentation. We will discuss what issues we have e ncountered while doing so and how we solved them.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/olav-bunte-formal-verific ation-of-oil-component-specifications-using-mcrl2/ END:VEVENT BEGIN:VEVENT UID:20200610T1416Z-1591798619.0243-EO-723-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200610T134746Z LAST-MODIFIED:20200610T134746Z DTSTART;TZID=Europe/Amsterdam:20200611T124500 DTEND;TZID=Europe/Amsterdam:20200611T133000 SUMMARY: Jeroen Keiren: Extensible Proof Systems for Infinite-State Systems DESCRIPTION: In the early 90’s\, Bradfield & Stirling developed a sound and complete tableau method for proving that sets of states in infinite-state labelled transition systems satisfy formulas in the modal mu-calculus. Unfo rtunately\, the soundness proof they presented is not extensible in the fol lowing sense. First\, it is hard to add new modalities to the mu-calculus\, ...continue reading X-ALT-DESC;FMTTYPE=text/html:Legacy software is a prominent bottleneck in modern industry: it is hard and costly to maintain\, yet contains valuab le knowledge not available elsewhere. Since manual rejuvenation of such sys tems is costly and cumbersome\, meta-programming techniques can be used to decrease this effort. Generic transformations\, however\, do not take domai n knowledge into account and are therefore not satisfactory. In this talk\, I will discuss an ongoing project I'm conducting together with Philips Hea lthcare\, in which we perform a significant refactoring on a part of their codebase. I will highlight how we have broken up this refactoring in smalle r transformations\, and how we evaluate the quality of each intermediate st ep.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rodin-aarssen-assessing-t he-quality-of-ad-hoc-refactorings/ END:VEVENT BEGIN:VEVENT UID:20200506T2217Z-1588803469.7411-EO-712-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200506T220155Z LAST-MODIFIED:20200506T220155Z DTSTART;TZID=Europe/Amsterdam:20200507T124500 DTEND;TZID=Europe/Amsterdam:20200507T133000 SUMMARY: Rick Erkens: Optimising adaptive non-linear pattern matching autom ata DESCRIPTION: This presentation builds on a previous presentation by Maurice Laveaux that was titled ‘adaptive non-linear pattern matching automata’. P attern matching is a fundamental problem in many areas of computing science \, both practical and theoretical. Given a large finite set of first-order patterns and a term t\, we are interested in an efficient way to find all . ..continue reading X-ALT-DESC;FMTTYPE=text/html:This presentation builds on a previous pre sentation by Maurice Laveaux that was titled 'adaptive non-linear pattern m atching automata'.
Pattern matching is a fundamental problem in many areas of computing science\, both practical and theoretical. Given a large finite set of first-order patterns and a term t\, we are interested in an e fficient way to find all pattern matches for t. In the previous presentatio n we proposed adaptive non-linear pattern matching automata (ANPMA) as an e xtension of earlier work on adaptive pattern matching automata by Sekar et al. By preprocessing the given pattern set and constructing an automaton\, it is possible to solve the matching problem by only inspecting every posit ion of term t at most once. In a setting where the pattern set is fixed and the matching problem has to be decided many times\, it is crucial for perf ormance to have such an automaton. The method that we proposed is not optim al yet. The construction algorithm can yield states that are redundant. The se redundancies are due to the fact that function symbol checks can give in formation about consistency checks and vice versa. The relation between the se two kinds of checks is not clear yet. In this presentation I will talk a bout a solution that we have in mind to compute an optimal ANPMA on the fly . The work on ANPMA will be presented at FSCD 2020. The extension that I wi ll talk about next thursday is still a work in progress.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rick-erkens-optimising-ad aptive-non-linear-pattern-matching-automata/ END:VEVENT BEGIN:VEVENT UID:20200428T1140Z-1588074054.1776-EO-710-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200428T113326Z LAST-MODIFIED:20200428T113326Z DTSTART;TZID=Europe/Amsterdam:20200430T124500 DTEND;TZID=Europe/Amsterdam:20200430T133000 SUMMARY: Hans Zantema: Computation of complexity of automatic sequences DESCRIPTION: In an earlier talk the complexity of an automatic sequence was presented as the minimal size of an automaton describing the sequence. Mor e precisely\, the n-th element of the binary sequence is 1 if and only if t he binary representation of n is accepted by the automaton. On this notion some main properties were investigated. In ...continue reading X-ALT-DESC;FMTTYPE=text/html:In an earlier talk the complexity of an au tomatic sequence was presented as the minimal size of an automaton describi ng the sequence. More precisely\, the n-th element of the binary sequence i s 1 if and only if the binary representation of n is accepted by the automa ton. On this notion some main properties were investigated. In the meantime a paper on this work has been accepted for the LATA 2020 conference\, and was honored by the best paper award. After a recap of the main notions and properties\, in this talk the focus will be on how to compute the correspon ding minimal automaton automatically from a prefix of the sequence\, by mea ns of SMT solving\, and on criteria on how long this prefix should be.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/hans-zantema-computation- of-complexity-of-automatic-sequences/ END:VEVENT BEGIN:VEVENT UID:20200422T0825Z-1587543917.3972-EO-709-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200421T201805Z LAST-MODIFIED:20200421T201805Z DTSTART;TZID=Europe/Amsterdam:20200423T124500 DTEND;TZID=Europe/Amsterdam:20200423T133000 SUMMARY: Maurice Laveaux: Decompositional Minimization of Monolithic Proces ses DESCRIPTION: Title: Decompositional Minimization of Monolithic Processes Ab stract: Compositional minimization is another established technique to tack le the state space explosion problem. This technique attempts to use the pa rallel processes defined in the high-level specification to obtain a reduce d state space w.r.t some behavioural equivalence immediately without explor ing the whole state space first. This technique would also ...continue read ing X-ALT-DESC;FMTTYPE=text/html:Title: Decompositional Minimization of Mon olithic Processes
Abstract: Compositional minimization is another est ablished technique to tackle the state space explosion problem. This techni que attempts to use the parallel processes defined in the high-level specif ication to obtain a reduced state space w.r.t some behavioural equivalence immediately without exploring the whole state space first. This technique w ould also be a useful addition for the mCRL2 toolset. However\, applying th is technique within the context of mCRL2 is hindered by the fact that we ty pically first transform the given specification into a monolithic process w here all parallel composition has been removed. Avoiding this transformatio n is undesirable\, because it facilitates various other static analysis tec hniques that might reduce the state space. Therefore\, we propose a method to decompose the resulting monolithic process\, where other simplification techniques could be applied first\, into a number of parallel processes suc h that the compositional minimization technique is applicable again.
CATEGORIES:Colloquium LOCATION:MS Teams GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/maurice-laveaux-decomposi tional-minimization-of-monolithic-processes/ END:VEVENT BEGIN:VEVENT UID:20200414T1406Z-1586873185.6511-EO-708-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200414T132541Z LAST-MODIFIED:20200414T132541Z DTSTART;TZID=Europe/Amsterdam:20200416T124500 DTEND;TZID=Europe/Amsterdam:20200416T133000 SUMMARY: Thomas Neele: Partial-Order Reduction for Parity Games with an App lication on Parameterised Boolean Equation Systems DESCRIPTION: Partial-order reduction (POR) is a well-established technique to combat the problem of state-space explosion. We propose POR techniques t hat are sound for parity games\, a well-established formalism for solving a variety of decision problems. As a consequence\, we obtain the first POR m ethod that is sound for model checking for the full modal mu-calculus. Our ...continue reading X-ALT-DESC;FMTTYPE=text/html:Partial-order reduction (POR) is a well-es tablished technique to combat the problem of state-space explosion. We prop ose POR techniques that are sound for parity games\, a well-established for malism for solving a variety of decision problems. As a consequence\, we ob tain the first POR method that is sound for model checking for the full mod al mu-calculus. Our technique is applied to\, and implemented for the fixed point logic called parameterised Boolean equation systems\, which provides a high-level representation of parity games. Experiments indicate that sub stantial reductions can be achieved.
CATEGORIES:Colloquium LOCATION:Zoom GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/thomas-neele-partial-orde r-reduction-for-parity-games-with-an-application-on-parameterised-boolean-e quation-systems/ END:VEVENT BEGIN:VEVENT UID:20200414T1406Z-1586873185.6578-EO-707-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200414T132104Z LAST-MODIFIED:20200414T132104Z DTSTART;TZID=Europe/Amsterdam:20200409T124500 DTEND;TZID=Europe/Amsterdam:20200409T133000 SUMMARY: Freek Verbeek: Formal Proofs of Return Address Integrity DESCRIPTION: We present a methodology for generating a characterization of the memory used by an assembly program\, as well as a formal proof that the assembly is bounded to the generated memory regions. A formal proof of mem ory usage is required for compositional reasoning over assembly programs. M oreover\, it can be used to prove low-level security ...continue reading X-ALT-DESC;FMTTYPE=text/html:We present a methodology for generating a characterization of the memory used by an assembly program\, as well as a f ormal proof that the assembly is bounded to the generated memory regions. A formal proof of memory usage is required for compositional reasoning over assembly programs. Moreover\, it can be used to prove low-level security pr operties\, such as integrity of the return address of a function. Our verif ication method is based on interactive theorem proving\, but provides autom ation by generating pre- and postconditions\, invariants\, control-flow\, a nd assumptions on memory layout. As a case study\, three binaries of the Xe n hypervisor are disassembled. These binaries are the result of a complex b uild-chain compiling production code\, and contain various complex and nest ed loops\, large and compound data structures\, and functions with over 100 basic blocks. The methodology has been successfully applied to 251 functio ns\, covering 12\,252 assembly instructions.
CATEGORIES:Colloquium LOCATION:Zoom GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/freek-verbeek-formal-proo fs-of-return-address-integrity/ END:VEVENT BEGIN:VEVENT UID:20200414T1406Z-1586873185.6607-EO-706-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200414T131644Z LAST-MODIFIED:20200414T131733Z DTSTART;TZID=Europe/Amsterdam:20200402T124500 DTEND;TZID=Europe/Amsterdam:20200402T133000 SUMMARY: Bas Luttik: Supporting railway infrastructure managers with formal models and analyses DESCRIPTION: In this talk\, I will discuss our recent experiences with usin g the mCRL2 toolset – which has a process-algebra based modelling language\ , a modal mu-calculus-based property language\, and an explicit-state model checker – to support two major innovation activities from railway infrastr ucture managers. First\, there is the EULYNX initiative of the European rai lway infrastructure managers. ...continue reading X-ALT-DESC;FMTTYPE=text/html:In this talk\, I will discuss our recent e xperiences with using the mCRL2 toolset - which has a process-algebra based modelling language\, a modal mu-calculus-based property language\, and an explicit-state model checker - to support two major innovation activities f rom railway infrastructure managers.
First\, there is the EULYNX init iative of the European railway infrastructure managers. The aim of EULYNX i s to standardise the interfaces between the interlocking and field elements (signals\, points\, level crossings)\; these interface standards are model led in SysML. In a project funded by the Dutch and German railway infrastru cture managers we are translating the SysML models to mCRL2 not only to for mally assess the quality of the standard by model checking\, but also to fa cilitate using them for model-based testing of compliance to the standard o f delivered components.
Second\, in collaboration with the Dutch rail way infrastructure manager ProRail we have formally modelled and analysed t he ERTMS Hybrid Level 3 principles. These principles facilitate subdividing track sections into virtual subsections\, in order to allow multiple train s simultaneously on the same track section\, thus increasing capacity. We h ave plans to support ProRail developers in their further elaboration of the design and implementation of ERTMS Hybrid Level 3.
CATEGORIES:Colloquium LOCATION:Zoom GEO:52.374540;4.897976 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/bas-luttik-supporting-rai lway-infrastructure-managers-with-formal-models-and-analyses/ END:VEVENT BEGIN:VEVENT UID:20200303T1238Z-1583239107.7407-EO-692-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200303T123149Z LAST-MODIFIED:20200303T123149Z DTSTART;TZID=Europe/Amsterdam:20200305T124500 DTEND;TZID=Europe/Amsterdam:20200305T133000 SUMMARY: Jan Martens: Resynchronizability of origin transducers DESCRIPTION: Origin semantics introduced by Bojańczyk is a fine grained sem antics for transducers\, that not only expresses the relation between input and output words\, but also includes a function that given an output posit ion returns the input position where it was produced: the origin. In this t alk we’ll discuss resynchronizations\, a tool to relax the notion ...contin ue reading X-ALT-DESC;FMTTYPE=text/html:
Origin semantics introduced by Bojańczyk i s a fine grained semantics for transducers\, that not only expresses the re lation between input and output words\, but also includes a function that g iven an output position returns the input position where it was produced: t he origin. In this talk we'll discuss resynchronizations\, a tool to relax the notion of origin while maintaining decidable properties. The talk speci fically focusses on the notion of containment up-to some unknown resynchron ization. This relation forms a pre-order strictly in between classical and origin containment. We show this containment is undecidable\, which was a r ecent open problem. This is done using a notion of limited traversal which syntactically classifies the non-existence of a resynchronization.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-martens-resynchroniza bility-of-origin-transducers/ END:VEVENT BEGIN:VEVENT UID:20200218T2105Z-1582059933.0252-EO-687-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200218T152034Z LAST-MODIFIED:20200218T152034Z DTSTART;TZID=Europe/Amsterdam:20200220T124500 DTEND;TZID=Europe/Amsterdam:20200220T133000 SUMMARY: Ferry Timmers: Maintaining strongly connected components efficient ly\, under edge deletions DESCRIPTION: Assume we have a directed graph. We can find the strongly conn ected components in this graph using for instance Tarjan’s algorithm. Let’s say we remove an edge from the graph. Can we recalculate the SCCs? A (rece nt) paper by Bernstein\, Probst et.al. presents a way tot do this efficient ly. I’ll roughly discuss my experiences with ...continue reading X-ALT-DESC;FMTTYPE=text/html:Assume we have a directed graph. We can fi nd the strongly connected components in this graph using for instance Tarja n's algorithm. Let's say we remove an edge from the graph. Can we recalcula te the SCCs? A (recent) paper by Bernstein\, Probst et.al. presents a way t ot do this efficiently. I'll roughly discuss my experiences with the algori thm and then focus on ES-trees\, a clever data structure the algorithm exte nded.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/ferry-timmers-maintaining -strongly-connected-components-efficiently-under-edge-deletions/ END:VEVENT BEGIN:VEVENT UID:20200219T1954Z-1582142088.944-EO-689-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200219T190328Z LAST-MODIFIED:20200219T190328Z DTSTART;TZID=Europe/Amsterdam:20200219T143000 DTEND;TZID=Europe/Amsterdam:20200219T151500 SUMMARY: Elbert van de Put: Ant Colony Optimization for Model Checking DESCRIPTION: Ant Colony Optimization is an optimization algorithm that is i nspired by the foraging behavior of ants. In this thesis I have applied Ant Colony Optimization to problems that are generated from the model checking problem\, to Boolean equation systems and to parity games. The results of this research are mixed but we have discovered approaches ...continue readi ng X-ALT-DESC;FMTTYPE=text/html:Ant Colony Optimization is an optimization algorithm that is inspired by the foraging behavior of ants. In this thesi s I have applied Ant Colony Optimization to problems that are generated fro m the model checking problem\, to Boolean equation systems and to parity ga mes. The results of this research are mixed but we have discovered approach es where further research seems promising.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 2 GEO:51.447551;5.487453 ORGANIZER;CN="Erik de Vink":MAILTO:e.p.d.vink@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/elbert-van-de-put-ant-col ony-optimization-for-model-checking/ END:VEVENT BEGIN:VEVENT UID:20200211T1329Z-1581427763.0372-EO-686-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200211T122925Z LAST-MODIFIED:20220203T111245Z DTSTART;TZID=Europe/Amsterdam:20200213T124500 DTEND;TZID=Europe/Amsterdam:20200213T133000 SUMMARY: Mark Bouwman: Verification and model-based testing of a railway po int/switch using mCRL2 DESCRIPTION: The EULYNX initiative is a collaborative effort of more than t en European railway infrastructure managers to standardise signalling inter faces. Within EULYNX\, FormaSig aims to use formal methods to analyse the c orrectness of the standard. We have recently concluded a case study in whic h we translate the EULYNX Point interface from SysML to mCRL2. The resultin g ...continue reading X-ALT-DESC;FMTTYPE=text/html:The EULYNX initiative is a collaborative e ffort of more than ten European railway infrastructure managers to standard ise signalling interfaces. Within EULYNX\, FormaSig aims to use formal methods to analyse the correctness of the standard. We have recently concluded a case study in which we trans late the EULYNX Point interface from SysML to mCRL2. The resulting mCRL2 mo del is subsequently used to contribute to the quality of the standard by ve rifying whether important safety requirements hold for the model. Test case s have been automatically derived from the same mCRL2 model. They have been executed\, effectively checking whether an actual Point implementation con forms to the EULYNX standard. This talk will discuss the case study itself and the lessons learned concerning formalising SysML and applying formal me thods in the railway domain.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/mark-bouwman-verification -and-model-based-testing-of-a-railway-point-switch-using-mcrl2/ END:VEVENT BEGIN:VEVENT UID:20200204T1113Z-1580814815.9488-EO-684-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200204T103823Z LAST-MODIFIED:20200204T103823Z DTSTART;TZID=Europe/Amsterdam:20200206T124500 DTEND;TZID=Europe/Amsterdam:20200206T133000 SUMMARY: Jan Bergstra: From computability theory to computer science and ba ck DESCRIPTION: I will discuss the sequence of themes which I have been resear ching since 1973\, and try to highlight how research questions have come\, and gone and reappeared. I will provide concise phrasings of leading princi ples which I have made use of in successive stages\, and how and why these have changed. Finally I will discuss ...continue reading X-ALT-DESC;FMTTYPE=text/html:I will discuss the sequence of themes whic h I have been researching since 1973\, and try to highlight how research qu estions have come\, and gone and reappeared. I will provide concise phrasin gs of leading principles which I have made use of in successive stages\, an d how and why these have changed. Finally I will discuss how the context of research changes (for the better) upon my "emeritaat"\, now 3 years ago. p> CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-bergstra-from-computa bility-theory-to-computer-science-and-back/ END:VEVENT BEGIN:VEVENT UID:20200121T1541Z-1579621268.5902-EO-678-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200121T141151Z LAST-MODIFIED:20200121T141151Z DTSTART;TZID=Europe/Amsterdam:20200123T124500 DTEND;TZID=Europe/Amsterdam:20200123T133000 SUMMARY: Jeroen Keiren: The basics of hardware verification using xMAS DESCRIPTION: xMAS is a graphical specification language that allows to desc ribe hardware components such as communication fabrics by composing microar chitectural primitives such as queues. Hardware components described using xMAS are sufficiently precise to allow formal verification\, however\, the explicit state space of such components suffers from the state space explos ion problem. Therefore\, other verification techniques have ...continue rea ding X-ALT-DESC;FMTTYPE=text/html:
xMAS is a graphical specification language that allows to describe hardware components such as communication fabrics by composing microarchitectural primitives such as queues. Hardware compone nts described using xMAS are sufficiently precise to allow formal verificat ion\, however\, the explicit state space of such components suffers from th e state space explosion problem. Therefore\, other verification techniques have been described in the literature. In particular\, sound (but not compl ete) approaches have been devised that translate the lack of liveness (some times referred to as the presence of a local deadlock) to SAT or SMT. In th is talk I will discuss the main aspects of the semantics of xMAS networks\, and I will discuss my understanding of some of the existing verification a pproaches.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jeroen-keiren-the-basics- of-hardware-verification-using-xmas/ END:VEVENT BEGIN:VEVENT UID:20200115T1203Z-1579089786.7184-EO-677-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200115T105130Z LAST-MODIFIED:20200115T105233Z DTSTART;TZID=Europe/Amsterdam:20200116T124500 DTEND;TZID=Europe/Amsterdam:20200116T133000 SUMMARY: Rodin Aarssen: High-Fidelity Metaprogramming with Separator Syntax Trees DESCRIPTION: Many metaprogramming tasks\, such as refactorings\, automated bug fixing\, or large-scale software renovation\, require high-fidelity sou rce code transformations – transformations which preserve comments and layo ut as much as possible. Abstract syntax trees (ASTs) typically abstract fro m such details\, and hence would require pretty printing\, destroying the o riginal program layout. Concrete syntax trees (CSTs) preserve all layout in formation\, ...continue reading X-ALT-DESC;FMTTYPE=text/html:Many metaprogramming tasks\, such as refac torings\, automated bug fixing\, or large-scale software renovation\, requi re high-fidelity source code transformations – transformations which preser ve comments and layout as much as possible. Abstract syntax trees (ASTs) ty pically abstract from such details\, and hence would require pretty printin g\, destroying the original program layout. Concrete syntax trees (CSTs) pr eserve all layout information\, but transformation systems or parsers that support CSTs are rare and can be cumbersome to use. In this talk I will pre sent separator syntax trees (SSTs)\, a lightweight syntax tree format\, tha t sits between AST and CSTs\, in terms of the amount of information they pr eserve. SSTs extend ASTs by recording textual layout information separating AST nodes. This information can be used to reconstruct the textual code af ter parsing\, but can largely be ignored when implementing high-fidelity tr ansformations.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rodin-aarssen-high-fideli ty-metaprogramming-with-separator-syntax-trees/ END:VEVENT BEGIN:VEVENT UID:20200107T1320Z-1578403252.1452-EO-675-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20200107T130837Z LAST-MODIFIED:20200107T130837Z DTSTART;TZID=Europe/Amsterdam:20200109T124500 DTEND;TZID=Europe/Amsterdam:20200109T133000 SUMMARY: Wieger Wesselink: A Dafny proof of Tarjan’s strongly connected com ponents algorithm DESCRIPTION: The mCRL2 toolset uses Tarjan’s strongly connected components algorithm at several places. To gain experience with automated provers\, we have made a proof of this algorithm using Dafny\, an automatic program ver ifier. In this talk I will explain our solution\, and tell something about our experiences with Dafny. This is joint work with Kees Huizing. X-ALT-DESC;FMTTYPE=text/html:The mCRL2 toolset uses Tarjan's strongly c onnected components algorithm at several places. To gain experience with au tomated provers\, we have made a proof of this algorithm using Dafny\, an a utomatic program verifier. In this talk I will explain our solution\, and t ell something about our experiences with Dafny.
This is joint work wi th Kees Huizing.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/wieger-wesselink-a-dafny- proof-of-tarjans-strongly-connected-components-algorithm/ END:VEVENT BEGIN:VEVENT UID:20191217T0925Z-1576574716.7118-EO-654-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20191217T091544Z LAST-MODIFIED:20191217T091544Z DTSTART;TZID=Europe/Amsterdam:20191219T124500 DTEND;TZID=Europe/Amsterdam:20191219T133000 SUMMARY: Rick Erkens: Abstraction for Bisimulation up to Context DESCRIPTION: Abstract: Roughly one year ago I gave a colloquium talk about up-to techniques for branching bisimilarity. In particular we discussed the soundness of branching bisimilarity up to context for CCS with guarded sum s. In combination with other up-to techniques this result allows for more f easible proofs when attempting to prove two CCS terms bisimilar. We added . ..continue reading X-ALT-DESC;FMTTYPE=text/html:Abstract:
Roughly one year ago I gav e a colloquium talk about up-to techniques for branching bisimilarity. In p articular we discussed the soundness of branching bisimilarity up to contex t for CCS with guarded sums. In combination with other up-to techniques thi s result allows for more feasible proofs when attempting to prove two CCS t erms bisimilar. We added two levels of abstraction to this result\, yieldin g the same result for a larger class of process algebras and allowing for e asier instantiation to other notions of bisimilarity that abstract from sil ent transitions. The first level involves Bloom's cool congruence formats. The second level involves a bialgebraic approach\, which will be the main f ocus of this talk.
Based on joint work with Jurriaan Rot and Bas Lutt ik.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rick-erkens-abstraction-f or-bisimulation-up-to-context/ END:VEVENT BEGIN:VEVENT UID:20191211T0921Z-1576056095.4876-EO-653-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20191211T091734Z LAST-MODIFIED:20191211T091734Z DTSTART;TZID=Europe/Amsterdam:20191212T124500 DTEND;TZID=Europe/Amsterdam:20191212T133000 SUMMARY: Tim Willemse: Family-Based SPL Model Checking Using Parity Games w ith Variability DESCRIPTION: We propose efficient family-based Software Product Lines (SPL) model checking based on variability parity games. These extend parity gam es with conditional edges labelled with (feature) configurations. We valida te our contribution by experiments on SPL benchmark models\, which demonstr ate that a novel family-based algorithm to collectively solve variability p arity games\, using symbolic representations of the configuration ...contin ue reading X-ALT-DESC;FMTTYPE=text/html:OIL is a domain-specific language under de velopment at Océ for specifying\, analysing\, and implementing software com ponents. OIL is to have IDE support\, transformations to formal modelling l anguages for requirement verification\, and code generation towards general -purpose languages such as C++. Model-based testing is an approach to test whether the behaviour of an implementation conforms to the behaviour descri bed in a formal model. A notable benefit of this approach is the automated test derivation and execution. This master thesis presents a C++ code gener ator for OIL\, implemented in the Spoofax language workbench. The correctne ss of the code generator is validated by means of model-based testing\, imp lemented in JTorX.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/mark-frenken-code-generat ion-and-model-based-testing-in-context-of-oil/ END:VEVENT BEGIN:VEVENT UID:20191204T1338Z-1575466733.2534-EO-649-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20191204T124753Z LAST-MODIFIED:20191204T124753Z DTSTART;TZID=Europe/Amsterdam:20191205T124500 DTEND;TZID=Europe/Amsterdam:20191205T133000 SUMMARY: Olav Bunte: Validity of OIL component specifications DESCRIPTION: With OIL one can model the desired behaviour of a system with a protocol specification and the actual behaviour of a component with a com ponent specification. To choose what events to execute\, an OIL component u ses a scheduler with run-to-completion semantics. However\, to avoid undesi red behaviour concerning this scheduler\, we need to put some validity ...c ontinue reading X-ALT-DESC;FMTTYPE=text/html:With OIL one can model the desired behavio ur of a system with a protocol specification and the actual behaviour of a component with a component specification. To choose what events to execute\ , an OIL component uses a scheduler with run-to-completion semantics. Howev er\, to avoid undesired behaviour concerning this scheduler\, we need to pu t some validity requirements on an OIL component specification. In this pre sentation we show what requirements these are and how they can be checked o n an OIL component specification.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/olav-bunte-validity-of-oi l-component-specifications/ END:VEVENT BEGIN:VEVENT UID:20191126T1058Z-1574765935.9939-EO-641-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20191126T104250Z LAST-MODIFIED:20191126T104250Z DTSTART;TZID=Europe/Amsterdam:20191128T124500 DTEND;TZID=Europe/Amsterdam:20191128T133000 SUMMARY: Maurice Laveaux: Adaptive Non-linear Pattern Match Automata DESCRIPTION: Efficient pattern matching is fundamental for practical term r ewrite engines. By preprocessing the given patterns into a finite determini stic automaton the matching patterns can be decided in a single traversal o f the relevant parts of the input term. Most of the existing automaton base d techniques are restricted to linear patterns\, where each variable occurs at ...continue reading X-ALT-DESC;FMTTYPE=text/html:Efficient pattern matching is fundamental for practical term rewrite engines. By preprocessing the given patterns int o a finite deterministic automaton the matching patterns can be decided in a single traversal of the relevant parts of the input term. Most of the exi sting automaton based techniques are restricted to linear patterns\, where each variable occurs at most once\, and require an additional post-processi ng step to check variable consistency. However\, interleaving the variable consistency and pattern matching phases can reduce the number of steps requ ired to find a match. Therefore\, we introduce consistency automata to dete rmine variable consistency using an optimal number of comparisons. These co nsistency automata are then combined with existing so-called adaptive patte rn match automata as introduced by Sekar et al. We show that the resulting automata can solve the pattern matching problem correctly\, and more effici ently then existing approaches.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/maurice-laveaux-adaptive- non-linear-pattern-match-automata/ END:VEVENT BEGIN:VEVENT UID:20191118T1425Z-1574087113.204-EO-623-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20191118T142420Z LAST-MODIFIED:20191118T142420Z DTSTART;TZID=Europe/Amsterdam:20191121T124500 DTEND;TZID=Europe/Amsterdam:20191121T133000 SUMMARY: Thomas Neele: The Inconsistent Labelling Problem of Stutter-Preser ving Partial-Order Reduction DESCRIPTION: In model checking\, partial-order reduction (POR) is an effect ive technique to reduce the size of the state space. Stubborn sets are an e stablished variant of POR and have seen many applications over the past 31 years. One of the early works on stubborn sets shows that a combination of several conditions on the reduction is ...continue reading X-ALT-DESC;FMTTYPE=text/html:In model checking\, partial-order reductio n (POR) is an effective technique to reduce the size of the state space. St ubborn sets are an established variant of POR and have seen many applicatio ns over the past 31 years. One of the early works on stubborn sets shows th at a combination of several conditions on the reduction is sufficient to pr eserve stutter-trace equivalence\, making stubborn sets suitable for model checking of linear-time properties. In this paper\, we identify a flaw in t he reasoning and show with a counter-example that stutter-trace equivalence is not necessarily preserved. We propose a solution and analyse in which f ormalisms this problem may occur. The impact on practical implementations i s limited\, since they all compute a correct approximation of the theory. p> CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/thomas-neele-the-inconsis tent-labelling-problem-of-stutter-preserving-partial-order-reduction/ END:VEVENT BEGIN:VEVENT UID:20191104T1616Z-1572884195.6778-EO-621-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20191104T160518Z LAST-MODIFIED:20191104T160518Z DTSTART;TZID=Europe/Amsterdam:20191107T124500 DTEND;TZID=Europe/Amsterdam:20191107T133000 SUMMARY: Tom Verhoeff: From FP to OO DESCRIPTION: This work got started while teaching a functional programming course\, where I tried to make the abstract concept of F-(co-)algebras more concrete by stating that\, in Java terms\, these are just classes that imp lement a specific kind of generic interface. To make this even more concret e\, I tried to implement this idea\, together with the ...continue reading X-ALT-DESC;FMTTYPE=text/html:
This work got started while teaching a fun ctional programming course\, where I tried to make the abstract concept of F-(co-)algebras more concrete by stating that\, in Java terms\, these are j ust classes that implement a specific kind of generic interface. To make th is even more concrete\, I tried to implement this idea\, together with the corresponding (co-)inductive type\, that is\, (terminal/)initial F-(co)-alg ebra. Somewhat to my surprise this worked out quite nicely. I present a sys tematic way of translating inductive types with their catamorphisms (folds) from those types\, and co-inductive types with their anamorphisms (unfolds ) into those types\, expressed as functional programs (say in Haskell)\, in to elegant object-oriented code (say in Java). Even types that are inductiv e and co-inductive can be handled\, allowing for hylomorphisms (fold after unfold) that are lazy. It turns out that there is a nice way of matching di sjoint sums (co-products)\, argument patterns\, and F-(co-)algebras on the FP side with inheritance\, polymorphism+overriding+dynamic dispatch\, and g eneric interfaces on the OO side. What also struck me is that this translat ion provides a legitimate use case for inheritance\, which is often abused (where composition would be more appropriate). I will illustrate this appro ach through some examples (trees\, streams\, and lists)\, while introducing the relevant FP and OO concepts. This serves as a Rosetta stone involving the three formalisms of Haskell\, Category Theory\, and Java. Even if you k now only one of these\, you should be able to grasp the others. The OO impl ementation allows for some useful trade-offs: memory versus time\, safety a nd easy reasoning versus efficiency and hard reasoning (think: immutable ve rsus mutable). I am working on a FP library for Java based on these ideas. Finally (when time permits)\, I will reflect on this translation and compar e (qualities of) the input FP code with (those of) the resulting OO code\, and with (those of) traditional OO code for the same functionality. I touch upon such qualities as performance\, code size\, maintainability\, and fle xibility.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tom-verhoeff-from-fp-to-o o/ END:VEVENT BEGIN:VEVENT UID:20191031T0829Z-1572510543.9305-EO-619-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20191031T081055Z LAST-MODIFIED:20191031T081055Z DTSTART;TZID=Europe/Amsterdam:20191106T130000 DTEND;TZID=Europe/Amsterdam:20191106T134500 SUMMARY: Sjef van Loo: Verifying SPLs using parity games expressing variabi lity DESCRIPTION: SPL verification can be costly when all the software products of an SPL are verified independently. It is well known that parity games ca n be used to verify software products. We propose a generalization of parit y games\, named variability parity games (VPGs)\, that encode multiple pari ty games in a single game graph decorated with edge ...continue reading X-ALT-DESC;FMTTYPE=text/html:SPL verification can be costly when all th e software products of an SPL are verified independently. It is well known that parity games can be used to verify software products. We propose a gen eralization of parity games\, named variability parity games (VPGs)\, that encode multiple parity games in a single game graph decorated with edge lab els expressing variability between the parity games. We show that a VPG can be constructed from a modal μ-calculus formula and an FTS that models the behaviour of the different software products of an SPL. Solving the resulti ng VPG decides for which products in the SPL the formula is satisfied. We i ntroduce several algorithms to efficiently solve VPGs and exploit commonali ties between the different parity games encoded. We perform experiments on SPL models to demonstrate that the VPG algorithms indeed outperform indepen dently verifying every product in an SPL.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/sjef-van-loo-verifying-sp ls-using-parity-games-expressing-variability/ END:VEVENT BEGIN:VEVENT UID:20191023T1512Z-1571843552.4571-EO-615-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20191023T143150Z LAST-MODIFIED:20191023T143150Z DTSTART;TZID=Europe/Amsterdam:20191024T124500 DTEND;TZID=Europe/Amsterdam:20191024T133000 SUMMARY: Jan Friso Groote: Internally optimized decision diagrams DESCRIPTION: When trying to understand ZDDs\, an alternative to BDDs\, I wa s wondering whether a quite different view on representing boolean function s would be interesting. I called this different view Internally Optimized D ecision Diagrams (IDDs). As it stands it is in no way clear whether this al ternative view would be beneficial to represent large boolean functions\, e specially because I ...continue reading X-ALT-DESC;FMTTYPE=text/html:When trying to understand ZDDs\, an altern ative to BDDs\, I was wondering whether a quite different view on represent ing boolean functions would be interesting. I called this different view In ternally Optimized Decision Diagrams (IDDs). As it stands it is in no way c lear whether this alternative view would be beneficial to represent large b oolean functions\, especially because I have difficulties finding efficient algorithms for standard functions\, such as "and" and "or"\, but I will st ill explain the basic ideas.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-friso-groote-internal ly-optimized-decision-diagrams/ END:VEVENT BEGIN:VEVENT UID:20191007T1513Z-1570461220.9654-EO-608-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20191007T125627Z LAST-MODIFIED:20191007T125627Z DTSTART;TZID=Europe/Amsterdam:20191010T124500 DTEND;TZID=Europe/Amsterdam:20191010T133000 SUMMARY: Vincenzo Ciancia: Spatial Model Checking and Applications to Medic al Image Analysis DESCRIPTION: Spatial aspects of computation are prominent in Computer Scien ce\, especially when dealing with systems distributed in physical space or with image data acquired from various sources. However\, formal verificatio n techniques are usually concerned with temporal properties and do not expl icitly handle spatial information. Our work stems from the topological inte rpretation of modal logics\, the so-called ...continue reading X-ALT-DESC;FMTTYPE=text/html:Spatial aspects of computation are promine
nt in Computer Science\, especially
when dealing with systems distribu
ted in physical space or with image data
acquired from various sources
. However\, formal verification techniques are
usually concerned with
temporal properties and do not explicitly handle spatial
information.<
/p>
Our work stems from the topological interpretation of modal logics\,
the
so-called Spatial Logics. We present a topology-based approach to
model checking
for spatial and spatio-temporal properties. Our results
include theoretical
developments in the more general setting of Cech
closure spaces\, a study of a
"collective" variant\, that has been com
pared to region calculi in recent work\,
publicly available software t
ools\, and some case studies in the setting of smart
transportation.
p>
In recent joint work with the University Hospital of Siena\, we have e
xplored the
application domain of automatic contouring in Medical Imag
ing\, introducing the
tool VoxLogicA\, which merges the state-of-the-a
rt imaging library ITK with the
unique combination of declarative spec
ification and optimised execution provided
by spatial model checking.
The analysis of an existing benchmark of medical
images for segmentati
on of brain tumours shows that simple VoxLogicA
specifications can rea
ch state-of-the-art accuracy\, competing with best-in-class
algorithms
based on machine learning\, with the advantage of explainability and
easy replicability.
References:
Vincenzo Ciancia\, Diego Latell
a\, Michele Loreti\, Mieke Massink:
Model Checking Spatial Logics for
Closure Spaces. Logical Methods in Computer Science 12(4) (2016)
https
://lmcs.episciences.org/2067
Gina Belmonte\, Vincenzo Ciancia\, Diego
Latella\, Mieke Massink:
VoxLogicA: A Spatial Model Checker for Decla
rative Image Analysis. TACAS (1) 2019: 281-298
https://link.springer.c
om/chapter/10.1007/978-3-030-17462-0_16
First order inductive theorem proving deal s with proving new equations based on a given set of equations. More specif ically\, we are interested in proving that the axioms logically imply the g oals. In this presentation I will discuss how we can automate these proofs by induction. The equations can either describe finite or infinite terms. I n this talk we are mainly interested in applying these ideas of induction o n infinite objects. The focus is on the most simple form of infinite object s: infinite sequences of some basic data type. Using a special operator tak e we can also apply our induction techniques on infinite data. Finally we d escribe how our techniques can be implemented in practice.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Hans Zantema":MAILTO:h.zantema@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/stan-roelofs-automaticall y-proving-equality-of-infinite-sequences/ END:VEVENT BEGIN:VEVENT UID:20190925T1434Z-1569422058.5098-EO-603-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190925T140236Z LAST-MODIFIED:20190925T140236Z DTSTART;TZID=Europe/Amsterdam:20190926T124500 DTEND;TZID=Europe/Amsterdam:20190926T133000 SUMMARY: Anton Wijs: Modular Indirect Push-button Formal Verification of Mu lti-threaded Code Generators DESCRIPTION: In model-driven development\, the automated generation of a mu lti-threaded program based on a model specifying the intended system behavi our is an important step. Verifying that such a generation step semanticall y preserves the specified functionality is hard. In related work\, code gen erators have been formally verified using theorem provers\, but this is ver y time-consuming work\, should ...continue reading X-ALT-DESC;FMTTYPE=text/html:In model-driven development\, the automate d generation of a multi-threaded program based on a model specifying the in tended system behaviour is an important step. Verifying that such a generat ion step semantically preserves the specified functionality is hard. In rel ated work\, code generators have been formally verified using theorem prove rs\, but this is very time-consuming work\, should be done by an expert in formal verification\, and is not easily adaptable to changes applied in the generator. In this paper\, we propose\, as an alternative\, a push-button approach\, combining equivalence checking and code verification with previo us results we obtained on the verification of generic code constructs. To i llustrate the approach\, we consider our SLCO framework\, which contains a generator of multi-threaded Java code. Although the technique can still onl y be applied to verify individual applications of the generator\, its push- button nature and efficiency in practice makes it very suitable for non-exp erts.
This is joint work with Maciej Wilkowski.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/anton-wijs-modular-indire ct-push-button-formal-verification-of-multi-threaded-code-generators/ END:VEVENT BEGIN:VEVENT UID:20190917T1431Z-1568730668.3326-EO-602-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190917T135637Z LAST-MODIFIED:20190917T135637Z DTSTART;TZID=Europe/Amsterdam:20190919T124500 DTEND;TZID=Europe/Amsterdam:20190919T133000 SUMMARY: Muhammad Osama: “SIGmA: GPU Accelerated Simplification of SAT Form ulas” DESCRIPTION: In this talk\, I will present SIGmA (SAT sImplification on GPU Architectures)\, a preprocessor to accelerate SAT solving that runs on NVI DIA GPU(s). We discuss the tool\, focusing on its full functionality\, incl uding new simplifications and multi-GPU support with load balancing mechani sm. SIGmA performs various types of simplification\, such as variable elimi nation\, subsumption elimination\, blocked ...continue reading X-ALT-DESC;FMTTYPE=text/html:In this talk\, I will present SIGmA (SAT s Implification on GPU Architectures)\, a preprocessor to accelerate SAT solv ing that runs on NVIDIA GPU(s). We discuss the tool\, focusing on its full functionality\, including new simplifications and multi-GPU support with lo ad balancing mechanism. SIGmA performs various types of simplification\, su ch as variable elimination\, subsumption elimination\, blocked clause elimi nation and hidden redundancy elimination. We study the effectiveness of our tool when applied prior to SAT solving. Overall\, for our large benchmark set of problems\, SIGmA enables MiniSat and Lingeling to solve many problem s in less time compared to applying the SatElite/Lingeling(integrated) prep rocessors.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/muhammad-osama-sigma-gpu- accelerated-simplification-of-sat-formulas/ END:VEVENT BEGIN:VEVENT UID:20190910T0952Z-1568109155.1282-EO-600-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190910T095221Z LAST-MODIFIED:20190910T095221Z DTSTART;TZID=Europe/Amsterdam:20190912T124500 DTEND;TZID=Europe/Amsterdam:20190912T133000 SUMMARY: Rob van Glabbeek: Is Speed-Independent Mutual Exclusion Implementa ble? DESCRIPTION: A mutual exclusion algorithm is called speed independent if it s correctness does not depend on the relative speed of the components. Famo us mutual exclusion protocols such as Dekker’s\, Peterson’s and Lamport’s b akery are meant to be speed independent. In this talk I argue that speed-in dependent mutual exclusion may not be implementable on standard hardware\, depending ...continue reading X-ALT-DESC;FMTTYPE=text/html:A mutual exclusion algorithm is called spe ed independent if its correctness does not depend on the relative speed of the components. Famous mutual exclusion protocols such as Dekker's\, Peters on's and Lamport's bakery are meant to be speed independent. In this talk I argue that speed-independent mutual exclusion may not be implementable on standard hardware\, depending on how we believe reading and writing to a me mory location is really carried out. It can be implemented on electrical ci rcuits\, however. This builds on previous work showing that mutual exclusio n cannot be accurately modelled in standard process algebras.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rob-van-glabbeek-is-speed -independent-mutual-exclusion-implementable/ END:VEVENT BEGIN:VEVENT UID:20190906T1128Z-1567769339.6213-EO-599-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190906T112229Z LAST-MODIFIED:20190910T142059Z DTSTART;TZID=Europe/Amsterdam:20190905T124500 DTEND;TZID=Europe/Amsterdam:20190905T133000 SUMMARY: David N. Jansen: Branching bisimulation – current status DESCRIPTION: Branching bisimulation is a behavioural equivalence relation o n labelled transition systems that takes internal actions into account. Whi le it is slightly coarser than weak bisimulation\, it has the advantage tha t there is a unique branching bisimulation equivalence quotient that can be found efficiently. With m the number of transitions and n the number of st ates\, ...continue reading X-ALT-DESC;FMTTYPE=text/html:Branching bisimulation is a behavioural eq uivalence relation on labelled transition systems that takes internal actio ns into account. While it is slightly coarser than weak bisimulation\, it h as the advantage that there is a unique branching bisimulation equivalence quotient that can be found efficiently. With m the number of transitions an d n the number of states\, the classic O(mn) algorithm has recently been re placed by an O(m log n) algorithm\, which is unfortunately rather complex. This paper provides a much more straightforward O(m log n) algorithm. Bench marks show that in practice this new algorithm is faster and more memory ef ficient than its predecessors.
(Joint work with Jan Friso Groote\, Je roen Keiren and Anton Wijs)
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/david-n-jansen-branching- bisimulation-current-status/ END:VEVENT BEGIN:VEVENT UID:20190815T1346Z-1565876794.6244-EO-590-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190815T133806Z LAST-MODIFIED:20190815T154213Z DTSTART;TZID=Europe/Amsterdam:20190826T150000 DTEND;TZID=Europe/Amsterdam:20190826T153000 SUMMARY: Msc presentation Johri van Eerd DESCRIPTION: Johri van Eerd will present his master thesis research on Mond ay August 26 at 15:00 in Atlas 2.215. His presentation addresses the questi on on whether term rewriting on GPUs is competitive compared to term rewrit ing on CPUs. X-ALT-DESC;FMTTYPE=text/html:Johri van Eerd will present his master the sis research on Monday August 26 at 15:00 in Atlas 2.215. His presentation addresses the question on whether term rewriting on GPUs is competitive com pared to term rewriting on CPUs.
CATEGORIES:MSc Defence LOCATION:Atlas 2.215 GEO:52.374540;4.897976 ORGANIZER;CN="Jan Friso Groote":MAILTO:j.f.groote@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/msc-presentation-johri-va n-eerd/ END:VEVENT BEGIN:VEVENT UID:20190812T1042Z-1565606570.2564-EO-587-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190812T091024Z LAST-MODIFIED:20190812T091024Z DTSTART;TZID=Europe/Amsterdam:20190812T140000 DTEND;TZID=Europe/Amsterdam:20190812T144500 SUMMARY: Kevin Nogarede: An approachable language for formal requirements DESCRIPTION: Formal system verification is a mathematical technique for est ablishing whether a process meets certain design requirements. Typically\, such techniques require notation in academic languages which are difficult for engineers to write and interpret. We aim to develop a new DSL for forma lizing requirements that dramatically lowers the barrier of entry by introd ucing notation and concepts ...continue reading X-ALT-DESC;FMTTYPE=text/html:Formal system verification is a mathematic al technique for establishing whether a process meets certain design requir ements. Typically\, such techniques require notation in academic languages which are difficult for engineers to write and interpret. We aim to develop a new DSL for formalizing requirements that dramatically lowers the barrie r of entry by introducing notation and concepts that are intuitively unders tandable yet still amenable to automated verification. The applicability of the new language is assessed via user experience reports
CATEGORIES:MSc Defence LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/kevin-nogarede-an-approac hable-language-for-formal-requirements/ END:VEVENT BEGIN:VEVENT UID:20190702T1547Z-1562082443.3325-EO-585-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190702T132921Z LAST-MODIFIED:20190702T132921Z DTSTART;TZID=Europe/Amsterdam:20190704T124500 DTEND;TZID=Europe/Amsterdam:20190704T133000 SUMMARY: David N. Jansen: Revisiting Weak Simulation for Substochastic Mark ov Chains DESCRIPTION: The spectrum of branching-time relations for probabilistic sys tems has been investigated thoroughly by Baier\, Hermanns\, Katoen and Wolf (2003\, 2005)\, including weak simulation for systems involving substochas tic distributions. Weak simulation was proven to be sound w. r. t. the live ness fragment of the logic PCTL\\X \, and its completeness was conjectured. We revisit this result ...continue reading X-ALT-DESC;FMTTYPE=text/html:The spectrum of branching-time relations f or probabilistic systems has been investigated thoroughly by Baier\, Herman ns\, Katoen and Wolf (2003\, 2005)\, including weak simulation for systems involving substochastic distributions. Weak simulation was proven to be sou nd w. r. t. the liveness fragment of the logic PCTL\\X \, and its completen ess was conjectured. We revisit this result and show that soundness does no t hold in general\, but only for Markov chains without divergence. It is re futed for some systems with substochastic distributions. Moreover\, we prov ide a counterexample to completeness. In this paper\, we present a novel de finition that is sound for live PCTL\\X \, and a variant that is both sound and complete.
A joint work with Lei Song and Lijun Zhang.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/david-n-jansen-revisiting -weak-simulation-for-substochastic-markov-chains/ END:VEVENT BEGIN:VEVENT UID:20190625T1629Z-1561480153.017-EO-584-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190625T160324Z LAST-MODIFIED:20190625T160324Z DTSTART;TZID=Europe/Amsterdam:20190627T124500 DTEND;TZID=Europe/Amsterdam:20190627T133000 SUMMARY: Mark Bouwman: Formal Modelling and Verification of an Interlocking using mCRL2 DESCRIPTION: This paper presents an application of the formal modelling and model checking toolkit mCRL2 and the model-based testing tool JTorX in the signalling domain. The mCRL2 toolkit is used to formally model the behavio ur of a system at the core of signalling solutions: the interlocking. The m odel of the interlocking is validated through model-based testing. We ...co ntinue reading X-ALT-DESC;FMTTYPE=text/html:This paper presents an application of the formal modelling and model checking toolkit mCRL2 and the model-based testi ng tool JTorX in the signalling domain. The mCRL2 toolkit is used to formal ly model the behaviour of a system at the core of signalling solutions: the interlocking. The model of the interlocking is validated through model-bas ed testing. We use the mCRL2 toolkit to verify high-level safety properties of the interlocking software.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/mark-bouwman-formal-model ling-and-verification-of-an-interlocking-using-mcrl2/ END:VEVENT BEGIN:VEVENT UID:20190625T1046Z-1561459595.5976-EO-582-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190625T103930Z LAST-MODIFIED:20190625T103930Z DTSTART;TZID=Europe/Amsterdam:20190626T140000 DTEND;TZID=Europe/Amsterdam:20190626T143000 SUMMARY: Ruud Meeuws: Model Checking Supermodels Workbench with mCRL2 DESCRIPTION: At Sioux\, a model-driven development tool is created. It allo ws users to create a model and generate software for specific hardware plat forms. For this tool\, Sioux wants to incorporate model checking in order t o improve the correctness and safety of software. To see whether model chec king is viable for industrial models\, we want to explore ...continue readi ng X-ALT-DESC;FMTTYPE=text/html:At Sioux\, a model-driven development tool
is created. It allows users to create a model and generate software for sp
ecific hardware platforms. For this tool\, Sioux wants to incorporate model
checking in order to improve the correctness and safety of software. To se
e whether model checking is viable for industrial models\, we want to explo
re the possibilities in limiting behaviour such that model checking becomes
feasible. We specify several execution models and see the influence of
them on the efficiency and feasibility of model checking with mCRL2 for t
his development tool.
Abstract: In the field of automata learnin g\, similar to model-based testing\, a target machine is sent interrogative queries with the goal of uncovering its behaviour. The most common algorit hm for this in the industry is L*_Mealy\, a Mealy-machine-learning variant of Dana Angluin's 1987 L*. In its current implementation\, no means of filt ering queries is available. In 2010\, however\, Fides Aarts and Frits Vaand rager presented a formalism that specifies certain patterns of restriction on the learner: an interface automaton they called the Learning Purpose. To gether with it\, they devised and implemented framework\, including a Mealy machine translator\, to learn interface automata\, which are similar to I/ O automata\, where the learner is restricted or rather guided through the l earning purpose automaton. In this talk\, I am going to explain their learn ing framework and focus on the learning purpose with the purpose of examini ng its expressivity in specifying desired patterns in learning\, and examin ing the possibility of extending it.
CATEGORIES:Colloquium LOCATION:MF 6.132 GEO:51.447551;5.487453 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/omar-alzuhaibi-automata-l earning-with-a-purpose/ END:VEVENT BEGIN:VEVENT UID:20190612T2218Z-1560377918.0069-EO-578-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190612T195857Z LAST-MODIFIED:20190612T195857Z DTSTART;TZID=Europe/Amsterdam:20190613T124500 DTEND;TZID=Europe/Amsterdam:20190613T133000 SUMMARY: Anton Wijs: “Check Your Locks and Fences: Integrating Model Checki ng into Fence and Transaction Insertion Analysis” DESCRIPTION: When targeting modern parallel hardware architectures\, constr ucting correct and high-performing software is complex and time-consuming. In particular\, reorderings of memory accesses that violate intended sequen tially consistent behaviour are a major source of bugs. Applying synchronis ation mechanisms to repair these should be done sparingly\, as they negativ ely impact performance. In this work\, we propose a technique ...continue r eading X-ALT-DESC;FMTTYPE=text/html:When targeting modern parallel hardware ar chitectures\, constructing correct and high-performing software is complex and time-consuming. In particular\, reorderings of memory accesses that vio late intended sequentially consistent behaviour are a major source of bugs. Applying synchronisation mechanisms to repair these should be done sparing ly\, as they negatively impact performance.
In this work\, we propose a technique that combines model checking and static analysis to analyse sp ecifications of multi-threaded software. For a given weak memory architectu re\, it identifies a set of fences to be placed in the corresponding code p roduced by our code generator. This set guarantees that the resulting softw are will be sequentially consistent. In cases where fences are not sufficie nt\, our technique suggests where to use stronger mechanisms\, such as lock ing or the use of transactional memory. In comparison\, static analysis tec hniques are less precise\, and model checking techniques have worse scalabi lity.
This is joint work with Sander de Putter.
CATEGORIES:Colloquium LOCATION:MetaForum 3.141 GEO:51.447567;5.487443 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/anton-wijs-check-your-loc ks-and-fences-integrating-model-checking-into-fence-and-transaction-inserti on-analysis/ END:VEVENT BEGIN:VEVENT UID:20190605T1405Z-1559743559.3721-EO-568-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190605T134701Z LAST-MODIFIED:20190605T134701Z DTSTART;TZID=Europe/Amsterdam:20190606T124500 DTEND;TZID=Europe/Amsterdam:20190606T133000 SUMMARY: Jan Friso Groote: An O(m log n) algorithm for branching bisimulati on DESCRIPTION: The O(m log n) algorithm for branching bisimulation devised by Groote/Jansen/Keiren/Wijs was primarily directed towards Kripke structures and not to labelled transition systems. To verify branching bisimulation f or LTSs an explicit translation is made to Kripke structures. This means th at the complexity for LTSs actually is O(m (log n + log |Act|)) and in prac tice the memory requirements are ...continue reading X-ALT-DESC;FMTTYPE=text/html:The O(m log n) algorithm for branching bis imulation devised by Groote/Jansen/Keiren/Wijs was primarily directed towar ds Kripke structures and not to labelled transition systems. To verify bran ching bisimulation for LTSs an explicit translation is made to Kripke struc tures. This means that the complexity for LTSs actually is O(m (log n + log |Act|)) and in practice the memory requirements are very high (but still O (m+n)). Together with David Jansen a dedicated algorithm has been construct ed which only requires O(m log n) and practically the current algorithm als o outperforms existing algorithms especially in memory requirements.
This is very much work in progress.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-friso-groote-an-om-lo g-n-algorithm-for-branching-bisimulation/ END:VEVENT BEGIN:VEVENT UID:20190522T1234Z-1558528459.4829-EO-564-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190522T105847Z LAST-MODIFIED:20190522T105847Z DTSTART;TZID=Europe/Amsterdam:20190523T124500 DTEND;TZID=Europe/Amsterdam:20190523T133000 SUMMARY: Hans Zantema: Symbolic model checking and bounded model checking DESCRIPTION: We discuss how to solve reachability problems either in BDD ba sed symbolic model checking (by NuXMV) or by bounded model checking using S MT solving. We provide several examples\, mainly from the practical assignm ents of the course Automated Reasoning\, over the years. These include dead lock checking in hardware and infinite branching. X-ALT-DESC;FMTTYPE=text/html:In this talk I will discuss the tangle lea rning algorithm for parity games. That is\, I will explain the concept of a tangle\, explain their role in different parity game algorithms\, and demo nstrate how we can find tangles in parity games. Furthermore I will introdu ce distractions\, simple distractions and devious distractions\, and the ro le they seem to play in parity games and how different algorithms deal with these distractions.
CATEGORIES:Colloquium LOCATION:MF 3.144 GEO:51.447551;5.487453 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tom-van-dijk-distractions -and-tangles-in-parity-games/ END:VEVENT BEGIN:VEVENT UID:20190508T1215Z-1557317728.6934-EO-549-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190508T120407Z LAST-MODIFIED:20190508T120407Z DTSTART;TZID=Europe/Amsterdam:20190509T124500 DTEND;TZID=Europe/Amsterdam:20190509T133000 SUMMARY: Ferry Timmers: A complete axiomatisation for probabilistic trace e quivalence DESCRIPTION: Probabilistic labelled transition systems combine the expressi veness of Markov chains and processes. We are interested in finding a suita ble trace equivalence for such systems\, which becomes non-trival once non- determinism is introduced. For this search I have axiomatised one of the eq uivalences\, based on weighed traces\, which is a complete axiomatisation\, to gain a better understanding ...continue reading X-ALT-DESC;FMTTYPE=text/html:Probabilistic labelled transition systems combine the expressiveness of Markov chains and processes. We are intereste d in finding a suitable trace equivalence for such systems\, which becomes non-trival once non-determinism is introduced. For this search I have axiom atised one of the equivalences\, based on weighed traces\, which is a compl ete axiomatisation\, to gain a better understanding of its properties.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/ferry-timmers-a-complete- axiomatisation-for-probabilistic-trace-equivalence-2/ END:VEVENT BEGIN:VEVENT UID:20190501T1447Z-1556722040.8849-EO-547-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190501T135037Z LAST-MODIFIED:20190501T135037Z DTSTART;TZID=Europe/Amsterdam:20190502T124500 DTEND;TZID=Europe/Amsterdam:20190502T133000 SUMMARY: Jeroen Keiren: Tableaux for mu-calculus model checking of infinite state systems: a note on soundness DESCRIPTION: In 1992\, Bradfield and Stirling presented a tableaux-based ap proach to local model checking of infinite-state systems that is sound and complete. When trying to apply this approach to model-checking timed and hy brid automata against mu-calculus properties we observed that the approach is not practical. In this talk I will explain Bradfield and Stirling’s proo f system\, ...continue reading X-ALT-DESC;FMTTYPE=text/html:In 1992\, Bradfield and Stirling presented a tableaux-based approach to local model checking of infinite-state system s that is sound and complete. When trying to apply this approach to model-c hecking timed and hybrid automata against mu-calculus properties we observe d that the approach is not practical. In this talk I will explain Bradfield and Stirling’s proof system\, and illustrate its limitations. I will then show how\, at the cost of losing completeness\, the proof system can be mad e more practical. Unfortunately\, the modifications break the existing soun dness proof. I therefore propose an alternative approach to the soundness p roof that is more robust when varying with the proof rules and termination conditions.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jeroen-keiren-tableaux-fo r-mu-calculus-model-checking-of-infinite-state-systems-a-note-on-soundness/ END:VEVENT BEGIN:VEVENT UID:20190423T1129Z-1556018944.2851-EO-545-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190423T111553Z LAST-MODIFIED:20190423T111553Z DTSTART;TZID=Europe/Amsterdam:20190425T124500 DTEND;TZID=Europe/Amsterdam:20190425T133000 SUMMARY: Allan van Hulst: The quest for the mythical 57-regular Moore graph DESCRIPTION: Since 1966 it is known that only four regular undirected graph s having diameter 2 and girth 5 can exist. The construction of three of the se 5\,2-Moore graphs is known but it is still an open problem whether the r emaining candidate srg(3250\,57\,0\,1) exists at all. I have developed an a lgorithm that constructs a partial solution to ...continue reading X-ALT-DESC;FMTTYPE=text/html:Since 1966 it is known that only four regu lar undirected graphs having diameter 2 and girth 5 can exist. The construc tion of three of these 5\,2-Moore graphs is known but it is still an open p roblem whether the remaining candidate srg(3250\,57\,0\,1) exists at all. I have developed an algorithm that constructs a partial solution to this pro blem. In particular\, I was able to remove all 3-cycles and 2 of the 3 type s of 4-cycles from an initial 57-regular graph of order 3250. During this t alk\, I would like to tap into possible algorithmic insights of members of the group to try and see whether anyone has any ideas concerning removal of the remaining 4-cycles. I will also discuss a number of earlier findings i n previous research. Notably\, and contrary to the other 5\,2-Moore graphs\ , this graph cannot be vertex-transitive and the order of its automorphism group is quite small (at most 375).
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/allan-van-hulst-the-quest -for-the-mythical-57-regular-moore-graph/ END:VEVENT BEGIN:VEVENT UID:20190416T1333Z-1555421608.1679-EO-542-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190416T132821Z LAST-MODIFIED:20190416T132821Z DTSTART;TZID=Europe/Amsterdam:20190418T124500 DTEND;TZID=Europe/Amsterdam:20190418T133000 SUMMARY: Rodin Aarssen: Concrete Syntax with Black Box Parsers DESCRIPTION: Meta programming is the art of writing software that takes sou rce code as input for manipulation\, analysis or code generation. Many meta programming systems reason about abstract syntax trees representing this s ource code\, which requires intimate knowledge of the data type that descri bes the abstract syntax. Concrete syntax patterns allow meta programmers to create and perform matching on syntax trees using ...continue reading X-ALT-DESC;FMTTYPE=text/html:Meta programming is the art of writing sof tware that takes source code as input for manipulation\, analysis or code g eneration. Many meta programming systems reason about abstract syntax trees representing this source code\, which requires intimate knowledge of the d ata type that describes the abstract syntax. Concrete syntax patterns allow meta programmers to create and perform matching on syntax trees using the actual concrete syntax of the object language. However\, meta programming s ystems that support these concrete syntax patterns generally require a conc rete grammar of the object language\, written in their own formalism. Writi ng such a grammar is a daunting\, error-prone task\, especially for non-tri vial programming languages\, such as C++ and Java.
We present Concret ely\, a technique to augment meta programming systems with pluggable concre te syntax\, reusing external\, black box parsers. Concretely allows the met a programmer to use concrete syntax patterns in absence of a concrete gramm ar. Additionally\, we present Tympanic\, a DSL to declaratively map externa l parsers' AST structures to the internal data structures of the Rascal met a programming language. Algebraic data types (ADTs) for the abstract gramma r and marshalling code\, mapping the external parser's AST to the generated Rascal ADT\, are automatically generated from a Tympanic specification. Ty mpanic allows implementors of Concretely to solve the impedance mismatch pr oblem between Rascal's algebraic data types and object-oriented class hiera rchies in Java representing a grammar.
We show that for realistic pro gramming languages\, such as C++ and Java\, the effort of adding support fo r concrete syntax patterns with Concretely is in the order of dozens of sou rce lines of code (SLOC). Similarly\, we show that using Tympanic for gramm ar mapping yields a significant reduction in terms of SLOC\, compared to ma nually implementing the AST data types and marshalling code.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rodin-aarssen-concrete-sy ntax-with-black-box-parsers/ END:VEVENT BEGIN:VEVENT UID:20190408T1105Z-1554721542.9025-EO-536-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190408T090245Z LAST-MODIFIED:20190408T090245Z DTSTART;TZID=Europe/Amsterdam:20190411T124500 DTEND;TZID=Europe/Amsterdam:20190411T133000 SUMMARY: Wieger Wesselink: About the Design and Implementation of State Spa ce Exploration in the mCRL2 toolset DESCRIPTION: An important use case of the mCRL2 tool set is state space exp loration. An efficient and feature-rich implementation is available that wo rks quite well in practice. Unfortunately the design of this implementation is quite complicated\, and there is insufficient documentation. This makes it very hard to make any changes\, or to add new functionality. During the ...continue reading X-ALT-DESC;FMTTYPE=text/html:An important use case of the mCRL2 tool se t is state space exploration. An efficient and feature-rich implementation is available that works quite well in practice. Unfortunately the design of this implementation is quite complicated\, and there is insufficient docum entation. This makes it very hard to make any changes\, or to add new funct ionality. During the last couple of months I have implemented a completely new version of state space exploration. In this talk I will explain its des ign and implementation. The new design contains pseudo code descriptions of the underlying algorithms. I consider it very important to have pseudo cod e descriptions at the right level of abstraction\, and an implementation th at closely matches the pseudo code. Another aspect that I will discuss is h ow to keep the code for all different features separate.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/wieger-wesselink-about-th e-design-and-implementation-of-state-space-exploration-in-the-mcrl2-toolset / END:VEVENT BEGIN:VEVENT UID:20190402T1121Z-1554204098.5243-EO-534-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190402T111006Z LAST-MODIFIED:20190402T111006Z DTSTART;TZID=Europe/Amsterdam:20190404T124500 DTEND;TZID=Europe/Amsterdam:20190404T133000 SUMMARY: Rick Erkens: “Rewriting the Term Rewriter Part 1: Introduction and Matching” DESCRIPTION: In an mCRL2 specification the user can specify processes with data\, that can in turn be manipulated through the computational model of t erm rewriting. The term rewriter that the toolset uses now is reasonably fa st and yet the bottleneck for state space generation is often found in the rewriter. This calls for optimizing this part ...continue reading X-ALT-DESC;FMTTYPE=text/html:Parity games are two-player\, infinite dur ation games that can be used answer whether a property holds true of a syst em with a yes or no. In several application domains\, one is not only inter ested in a yes/no answer\, but in computing some/all values to parameters i n a system description for which a property holds true. A naive approach to this problem involves solving many parity games. In the hope of taking adv antage of structural commonalities among a collection of games\, thus avoid ing many unnecessary computations\, we introduce variability parity games a s a generalisation of parity games. We will briefly discuss an algorithm th at directly solves such variability parity games. As a motivation for\, and an application of the developed theory\, we suggest how it can be used in the setting of Software Product Lines for modal mu-calculus model checking for Featured Transition Systems. This is very much work in (slow) progress. ..
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tim-willemse-variability- parity-games/ END:VEVENT BEGIN:VEVENT UID:20190320T1319Z-1553087999.2203-EO-531-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190320T093442Z LAST-MODIFIED:20190320T093442Z DTSTART;TZID=Europe/Amsterdam:20190321T124500 DTEND;TZID=Europe/Amsterdam:20190321T133000 SUMMARY: Jurriaan Rot: Coalgebra Learning via Duality DESCRIPTION: Automata learning is a popular technique for inferring minimal automata through membership and equivalence queries. We generalise learnin g from automata to a large class of state-based systems\, using the theory of coalgebras. The approach relies on the use of logical formulas as tests\ , based on a dual adjunction between states and logical theories. This allo ws ...continue reading X-ALT-DESC;FMTTYPE=text/html:Automata learning is a popular technique f
or inferring minimal
automata through membership and equivalence queri
es. We generalise
learning from automata to a large class of state-bas
ed systems\, using
the theory of coalgebras. The approach relies on th
e use of logical
formulas as tests\, based on a dual adjunction betwee
n states and
logical theories. This allows us to learn\, e.g.\, labell
ed transition
systems.
Joint work with Clemens Kupke and Simone Barlocco.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jurriaan-rot-coalgebra-le arning-via-duality/ END:VEVENT BEGIN:VEVENT UID:20190313T1105Z-1552475152.8866-EO-529-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190313T103323Z LAST-MODIFIED:20190313T103323Z DTSTART;TZID=Europe/Amsterdam:20190314T124500 DTEND;TZID=Europe/Amsterdam:20190314T133000 SUMMARY: Olav Bunte: “Formalising the semantics of OIL: the first steps” DESCRIPTION: In the last years an increasing number of companies have shown interest in applying verification techniques in model based software engin eering and Océ is one of them. To apply such techniques it is necessary to have a formal semantics of the modelling language used\, which in our case is OIL (Océ Interaction Language). We show ...continue reading X-ALT-DESC;FMTTYPE=text/html:In the last years an increasing number of companies have shown interest in applying verification techniques in model based software engineering and Océ is one of them. To apply such techniques it is necessary to have a formal semantics of the modelling language used\ , which in our case is OIL (Océ Interaction Language). We show the basic an d important ingredients of OIL and give their semantics. We also show how t he formalisation of these semantics\, using a transformation to mCRL2\, is realised.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/olav-bunte-formalising-th e-semantics-of-oil-the-first-steps/ END:VEVENT BEGIN:VEVENT UID:20190226T1251Z-1551185463.4125-EO-518-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190226T123412Z LAST-MODIFIED:20190226T123412Z DTSTART;TZID=Europe/Amsterdam:20190228T124500 DTEND;TZID=Europe/Amsterdam:20190228T133000 SUMMARY: Maurice Laveaux: Correct and Efficient Antichain Algorithms for Re finement Checking DESCRIPTION: Refinement checking plays an important role in system verifica tion. This means that the correctness of the system is established by showi ng a refinement relation between two models\; one for the implementation an d one for the specification. Previously\, Wang et al. presented an algorith m based on antichains to efficiently decide stable failures refinement and failures-divergences refinement. ...continue reading X-ALT-DESC;FMTTYPE=text/html:Refinement checking plays an important rol e in system verification. This means that the correctness of the system is established by showing a refinement relation between two models\; one for t he implementation and one for the specification. Previously\, Wang et al. p resented an algorithm based on antichains to efficiently decide stable fail ures refinement and failures-divergences refinement. We have identified sev eral issues pertaining to the correctness and performance of these algorith ms and propose new\, correct\, antichain-based algorithms. Using a number o f experiments we show that our algorithms outperform the original ones in t erms of running time and memory usage.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/maurice-laveaux-correct-a nd-efficient-antichain-algorithms-for-refinement-checking/ END:VEVENT BEGIN:VEVENT UID:20190219T1431Z-1550586702.6719-EO-517-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190219T125317Z LAST-MODIFIED:20190219T125317Z DTSTART;TZID=Europe/Amsterdam:20190221T124500 DTEND;TZID=Europe/Amsterdam:20190221T133000 SUMMARY: Thomas Neele: Verifying System-Wide Properties of Industrial Compo nent-Based Software DESCRIPTION: Analytical Software Design (ASD) enables model-based developme nt of component software systems. Until now\, functional verification of AS D systems is only possible on a per-component basis. There is no functional verification engine for ASD itself\, so this verification relies on a tran slation of individual components to mCRL2\, a process-algebraic model check er. We show how to extend ...continue reading X-ALT-DESC;FMTTYPE=text/html:Analytical Software Design (ASD) enables m
odel-based development of
component software systems. Until now\, func
tional verification of ASD
systems is only possible on a per-component
basis. There is no
functional verification engine for ASD itself\, so
this verification
relies on a translation of individual components to
mCRL2\, a
process-algebraic model checker. We show how to extend the
ASD-mCRL2
translation to support multiple components in order to enabl
e checking
of system wide functional properties. With our extended tra
nslation\, we
perform a case-study on a newly developed industrial sys
tem consisting
of 26 communicating components. The results indicate th
at it is feasible
to model check functional properties on this scale.<
/p>
CATEGORIES:Colloquium
LOCATION:MetaForum MF 7.084
GEO:51.447532;5.487437
ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl
URL;VALUE=URI:https://fsa.win.tue.nl/events/event/thomas-neele-verifying-sy
stem-wide-properties-of-industrial-component-based-software/
END:VEVENT
BEGIN:VEVENT
UID:20190213T1449Z-1550069377.9378-EO-514-1@37.128.148.44
STATUS:CONFIRMED
DTSTAMP:20240319T075845Z
CREATED:20190213T143044Z
LAST-MODIFIED:20190213T143044Z
DTSTART;TZID=Europe/Amsterdam:20190214T124500
DTEND;TZID=Europe/Amsterdam:20190214T133000
SUMMARY: Alexander Fedotov: Fixing block and idle equations for FSMs in MaD
L (work in progress)
DESCRIPTION: The xMAS language introduced a convenient way of high-level mo
deling and verification of communication fabrics. For micro-architectural m
odels expressed in xMAS\, it was shown that the problem of proving liveness
could be transformed into a SAT-problem\, which is advantageous in terms o
f scalability. Later on\, Verbeek et al. proposed an approach to combine ba
sic xMAS ...continue reading
X-ALT-DESC;FMTTYPE=text/html:
The xMAS language introduced a convenient way of high-level modeling and verification of communication fabrics. For m icro-architectural models expressed in xMAS\, it was shown that the problem of proving liveness could be transformed into a SAT-problem\, which is adv antageous in terms of scalability. Later on\, Verbeek et al. proposed an ap proach to combine basic xMAS primitives with Finite State Machines. The ide a itself resulted in a more expressive\, yet scalable technique for modelin g and verification of microarchitectures. However\, we found a mistake in i t. In the current work\, we show that the technique introduced by Verbeek e t al. is not sound\, i.e.\, it fails to prove deadlock freedom. We propose a new method which addresses the problems of the previous approach.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/alexander-fedotov-fixing- block-and-idle-equations-for-fsms-in-madl-work-in-progress/ END:VEVENT BEGIN:VEVENT UID:20190206T0931Z-1549445506.2991-EO-512-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190206T082009Z LAST-MODIFIED:20190206T082009Z DTSTART;TZID=Europe/Amsterdam:20190207T124500 DTEND;TZID=Europe/Amsterdam:20190207T133000 SUMMARY: Erik de Vink: On a complete axiomatization of rooted branching pro babilistic bisimulation DESCRIPTION: We consider a process language featuring both nondeterministic and probabilistic choice with an operational semantics taking distribution s of processes as basic building blocks. For this langauge we propose a sou nd and complete axiomatization of rooted branching probabilistic bisimulati on. Exploiting the notion of a concrete process and building on the complet eness of strong probabilistic bisimulation\, a ...continue reading X-ALT-DESC;FMTTYPE=text/html:We consider a process language featuring b
oth nondeterministic and
probabilistic choice with an operational sema
ntics taking
distributions of processes as basic building blocks. For
this langauge
we propose a sound and complete axiomatization of rooted
branching
probabilistic bisimulation. Exploiting the notion of a conc
rete
process and building on the completeness of strong probabilistic<
br />bisimulation\, a completeness result for rooted branching probabilisti
c
bisimulation is obtained.
Joint work with Jan Friso Groote
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/erik-de-vink-on-a-complet e-axiomatization-of-rooted-branching-probabilistic-bisimulation/ END:VEVENT BEGIN:VEVENT UID:20190131T1923Z-1548962620.691-EO-506-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190131T144556Z LAST-MODIFIED:20190131T144556Z DTSTART;TZID=Europe/Amsterdam:20190205T110000 DTEND;TZID=Europe/Amsterdam:20190205T114500 SUMMARY: Mark Bouwman: A model-based test platform for rail signalling syst ems DESCRIPTION: As technology progresses\, newer\, and more complex\, solution s are employed to verify that rail signalling systems are safe. Formal meth ods provide ways to increase rigour in the verification process. This preci sion\, accompanied by the ongoing increase of computational power of comput ers\, also opens up ways to partially automate parts of the verification pr ocess. We present ...continue reading X-ALT-DESC;FMTTYPE=text/html:As technology progresses\, newer\, and mor e complex\, solutions are employed to verify that rail signalling systems a re safe. Formal methods provide ways to increase rigour in the verification process. This precision\, accompanied by the ongoing increase of computati onal power of computers\, also opens up ways to partially automate parts of the verification process. We present a case study an application of mCRL2 and the model-based testing tool JTorX in the signalling domain. The mCRL2 toolkit is used to formally model the behaviour of a system at the core of signalling solutions: the interlocking. The behaviour of the interlocking i s validated through model checking\, proving that relevant safety propertie s hold. Using JTorX\, the formal model is turned into the benchmark in an a utomated testing platform for interlocking software. A working setup with a ctual interlocking software on a pre-existing testing platform is presented \, though performance and stability remain an issue. The suitability of mCR L2 and JTorX in the signalling domain is evaluated and suggestions are give n for improvement and further research.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 3 GEO:51.447567;5.487443 ORGANIZER;CN="Bas Luttik":MAILTO:s.p.luttik@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/mark-bouwman-a-model-base d-test-platform-for-rail-signalling-systems/ END:VEVENT BEGIN:VEVENT UID:20190129T2122Z-1548796927.4215-EO-502-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190129T135403Z LAST-MODIFIED:20190129T135403Z DTSTART;TZID=Europe/Amsterdam:20190131T124500 DTEND;TZID=Europe/Amsterdam:20190131T131500 SUMMARY: Bas Luttik: Revised Semantics for Sequential Composition in the Pr esence of Intermediate Acceptance DESCRIPTION: For a smooth integration of classical automata theory and conc urrency theory\, one would like to consider process algebras including a co nstant denoting successful termination and binary operations for alternativ e and sequential composition. Using alternative composition and successful termination it is possible to express a notion of intermediate acceptance a s it occurs\, e.g.\, in (classical) finite ...continue reading X-ALT-DESC;FMTTYPE=text/html:For a smooth integration of classical auto mata theory and concurrency theory\, one would like to consider process alg ebras including a constant denoting successful termination and binary opera tions for alternative and sequential composition. Using alternative composi tion and successful termination it is possible to express a notion of inter mediate acceptance as it occurs\, e.g.\, in (classical) finite automata. Se quential composition is necessary for the process-theoretic counterpart of context-free grammars. A process algebra including said ingredients with th eir traditional semantics\, however\, also inherits the less desirable phen omenon of transparency: behaviour of the first component of a sequential co mposition may be skipped.
We present a revised semantics for sequenti al composition that retains the properties needed for a smooth integration of classical automata theory and concurrency theory\, but eliminates transp arency. For the resulting process algebra we have obtained axiomatisation a nd decidability results.
(based on joint work with Jos Baeten\, Astri d Belder and Fei Yang)
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/bas-luttik-revised-semant ics-for-sequential-composition-in-the-presence-of-intermediate-acceptance/ END:VEVENT BEGIN:VEVENT UID:20190115T1415Z-1547561732.4671-EO-490-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190115T063254Z LAST-MODIFIED:20190115T063254Z DTSTART;TZID=Europe/Amsterdam:20190117T124500 DTEND;TZID=Europe/Amsterdam:20190117T131500 SUMMARY: Hans Zantema: Equational reasoning\, induction\, and infinite data structures DESCRIPTION: Inductive theorem proving combines equational reasoning and in duction to prove properties on ground terms. We present some basics\, and s how how it also applies in data structures not having a unique representati on and in infinite data structures. X-ALT-DESC;FMTTYPE=text/html:Inductive theorem proving combines equatio nal reasoning and induction to prove properties on ground terms. We present some basics\, and show how it also applies in data structures not having a unique representation and in infinite data structures.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/hans-zantema-equational-r easoning-induction-and-infinite-data-structures/ END:VEVENT BEGIN:VEVENT UID:20190109T0230Z-1547001003.7325-EO-488-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20190108T152049Z LAST-MODIFIED:20190108T152049Z DTSTART;TZID=Europe/Amsterdam:20190110T124500 DTEND;TZID=Europe/Amsterdam:20190110T131500 SUMMARY: Sofie Haesaert: Robust Dynamic Programming for Temporal Logic Cont rol of Stochastic Systems DESCRIPTION: Discrete-time stochastic systems are an essential modelling to ol for many engineering systems. We consider stochastic control systems tha t are evolving over continuous spaces. For this class of models\, methods f or the formal verification and synthesis of control strategies are computat ionally hard and generally rely on the use of approximate abstractions. Bui lding on approximate abstractions\, we ...continue reading X-ALT-DESC;FMTTYPE=text/html:Discrete-time stochastic systems are an es sential modelling tool for many engineering systems. We consider stochastic control systems that are evolving over continuous spaces. For this class o f models\, methods for the formal verification and synthesis of control str ategies are computationally hard and generally rely on the use of approxima te abstractions. Building on approximate abstractions\, we compute control strategies with lower- and upper-bounds for satisfying unbounded temporal l ogic specifications. Firstly\, robust dynamic programming mappings over the abstract system are introduced to solve the control synthesis and verifica tion problem. These mappings yield a control strategy and a unique lower bo und on the satisfaction probability for temporal logic specifications that is robust to the incurred approximation errors. Secondly\, upper-bounds on the satisfaction probability are quantified\, and properties of the mapping s are analysed and discussed. Finally\, we show the implications of these r esults for linear stochastic dynamic systems with a continuous state space. This abstraction-based synthesis framework is shown to be able to handle i nfinite-horizon properties. Approximation errors expressed as deviations in the outputs of the models and as deviations in the probabilistic transitio ns are allowed and are quantified using approximate stochastic simulation r elations.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/sofie-haesaert-robust-dyn amic-programming-for-temporal-logic-control-of-stochastic-systems/ END:VEVENT BEGIN:VEVENT UID:20181219T0034Z-1545179699.8399-EO-486-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20181218T150540Z LAST-MODIFIED:20181218T150540Z DTSTART;TZID=Europe/Amsterdam:20181220T124500 DTEND;TZID=Europe/Amsterdam:20181220T131500 SUMMARY: Maciej Gazda: Logical characterisation of hybrid conformance DESCRIPTION: The notion of conformance provides a rigorous basis for testin g systems. In particular\, a notion of hybrid conformance is useful in esta blishing a formal model-based technique for cyber-physical systems. A logic al characterisation of conformance precisely specifies the set of formulae that are preserved and reflected by conformance. We present what is to our knowledge\, the ...continue reading X-ALT-DESC;FMTTYPE=text/html:The notion of conformance provides a rigor
ous basis for testing
systems. In particular\, a notion of hybrid conf
ormance is useful in
establishing a formal model-based technique for c
yber-physical
systems. A logical characterisation of conformance preci
sely specifies
the set of formulae that are preserved and reflected by
conformance.
We present what is to our knowledge\, the first characte
risation result
for an approximate notion of hybrid conformance. To th
is end\, we show
that the relaxation scheme used for preservation resu
lts in this
setting is not tight enough for providing a characterisati
on and
propose a tighter relaxation that we subsequently prove to be t
he
right one for characterising hybrid conformance.
In this talk\, first I give an overview of activities in our session type group in Imperial and how mCRL2 is used in our researches.
Go is a production-level statically typed programming language whose design features explicit message-passing primitives and lig htweight threads\, enabling (and encouraging) programmers to develop concur rent systems where components interact through communication more so than b y lock-based shared memory concurrency. Go can detect global deadlocks at r untime\, but does not provide any compile-time protection against all too c ommon communication mismatches and partial deadlocks.
In this work\, we present a static verification framework for liveness and safety in Go pr ograms\, able to detect communication errors and deadlocks by mCRL2. Our to olchain infers from a Go program a faithful representation of its communica tion patterns as behavioral types\, where the types are model checked for l iveness and safety.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/nobuko-yoshida-behavioura l-type-based-static-verification-framework-for-go/ END:VEVENT BEGIN:VEVENT UID:20181204T1553Z-1543938796.7268-EO-481-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20181204T140617Z LAST-MODIFIED:20181204T140617Z DTSTART;TZID=Europe/Amsterdam:20181206T124500 DTEND;TZID=Europe/Amsterdam:20181206T131500 SUMMARY: Muhammad Osama: SAT Solving using Ant Colony Optimization on GPUs DESCRIPTION: In this talk\, I will discuss how the Boolean Satisfiability ( SAT) problem can be solved using a parallel implementation of the Ant Colon y Optimization (ACO) algorithm for execution on the Graphics Processing Uni t (GPU). We propose a new efficient parallel strategy for the ACO algorithm executed entirely on the CUDA architecture\, and perform experiments to .. .continue reading X-ALT-DESC;FMTTYPE=text/html:In this talk\, I will discuss how the Bool ean Satisfiability (SAT) problem can be solved using a parallel implementat ion of the Ant Colony Optimization (ACO) algorithm for execution on the Gra phics Processing Unit (GPU). We propose a new efficient parallel strategy f or the ACO algorithm executed entirely on the CUDA architecture\, and perfo rm experiments to compare it with the best sequential version exists implem ented on CPU with incomplete approaches. We show how SAT problem can benefi t from the GPU solutions\, leading to significant improvements in speed-up even though keeping the quality of the solution. Our results shows that the new parallel implementation executes up to 21x faster compared to its sequ ential counterpart.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/muhammad-osama-sat-solvin g-using-ant-colony-optimization-on-gpus/ END:VEVENT BEGIN:VEVENT UID:20181128T1936Z-1543433796.8335-EO-480-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20181128T131041Z LAST-MODIFIED:20181128T131041Z DTSTART;TZID=Europe/Amsterdam:20181129T124500 DTEND;TZID=Europe/Amsterdam:20181129T131500 SUMMARY: Anton Wijs: The SLCO Framework for Verified\, Model-driven Constru ction of Component Software DESCRIPTION: I will present the Simple Language of Communicating Objects (\ \SLCO) framework\, which has resulted from our research on applying formal methods for correct and efficient model-driven development of multi-compone nt software. At the core is a domain specific language called SLCO that can be used to specify software behaviour. One of the features of the framewor k ...continue reading X-ALT-DESC;FMTTYPE=text/html:I will present the Simple Language of Comm unicating Objects (\\SLCO) framework\, which has resulted from our research on applying formal methods for correct and efficient model-driven developm ent of multi-component software. At the core is a domain specific language called SLCO that can be used to specify software behaviour. One of the feat ures of the framework is the formal verification of SLCO models\, which is provided via a translation to mCRL2. In this talk\, I will discuss the lang uage SLCO\, give an overview of the features of the framework\, and discuss our roadmap for the future.
CATEGORIES:Colloquium LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/anton-wijs-the-slco-frame work-for-verified-model-driven-construction-of-component-software/ END:VEVENT BEGIN:VEVENT UID:20181121T1807Z-1542823679.1154-EO-475-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20181121T093528Z LAST-MODIFIED:20181121T093528Z DTSTART;TZID=Europe/Amsterdam:20181122T124500 DTEND;TZID=Europe/Amsterdam:20181122T131500 SUMMARY: Jan Friso Groote: A game that I can win\, but don’t know how DESCRIPTION: After the PhD. defense of Maciej Gazda\, I walked back with on e of the committee members who explained me a game that he knew he could wi n\, but he would not be able to tell how. X-ALT-DESC;FMTTYPE=text/html:After the PhD. defense of Maciej Gazda\, I
walked back with one of the committee members
who explained me a game
that he knew he could win\, but he would not be able to tell how.
At ASML the ASOME data modelling language has been defined. Many system designers are defining models in this languag e\, and the question is whether there is an easy way to assess that these m odels are ok. This can be done using metrics. The question is which metrics are proper. In this MSc research various metrics from different sources ar e collected. The suitability of these metrics to indicate the complexity an d maintainability of the models is assessed by questionaires and qualitati ve interviews.
CATEGORIES:MSc Defence LOCATION:Metaforum MF3.122 GEO:52.374540;4.897976 ORGANIZER;CN="Jan Friso Groote":MAILTO:j.f.groote@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/hong-zhang-quality-metric s-for-asome-data-models/ END:VEVENT BEGIN:VEVENT UID:20181114T1259Z-1542200383.5658-EO-471-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20181114T101903Z LAST-MODIFIED:20181114T101903Z DTSTART;TZID=Europe/Amsterdam:20181115T124500 DTEND;TZID=Europe/Amsterdam:20181115T131500 SUMMARY: M. Laveaux\, T. Neele\, O. Bunte: The 201808.0 mCRL2 release DESCRIPTION: In this talk\, we will give an overview of the additions and i mprovements to mCRL2 from the last couple of years. First\, the mCRL2 langu age has been expanded with the possibility to specify probabilistic behavio ur. Second\, there are several new and improved implementations of behaviou ral relations: the GJKW algorithm for branching bisimulation with complexit y O(m ...continue reading X-ALT-DESC;FMTTYPE=text/html:In this talk\, we will give an overview of the additions and improvements to mCRL2 from the last couple of years. Fir st\, the mCRL2 language has been expanded with the possibility to specify p robabilistic behaviour. Second\, there are several new and improved impleme ntations of behavioural relations: the GJKW algorithm for branching bisimul ation with complexity O(m log n)\, several new preorders and an implementat ion of probabilistic branching bisimulation. Model checking with the modal mu-calculus has been expanded with the possibility to generate evidence in the form of a counter-example or witness. Furthermore\, there are some new techniques to support model checking of infinite-state systems. Finally\, t here is a new\, more intuitive tool for developing specifications. This too l is called mCRL2ide and it support visualisation of state spaces and also model checking of mu-calculus properties. Throughout the talk\, we will do several demos.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/m-laveaux-t-neele-o-bunte -the-201808-0-mcrl2-release/ END:VEVENT BEGIN:VEVENT UID:20181107T1041Z-1541587292.6296-EO-467-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20181107T101005Z LAST-MODIFIED:20181107T101005Z DTSTART;TZID=Europe/Amsterdam:20181108T124500 DTEND;TZID=Europe/Amsterdam:20181108T131500 SUMMARY: Rick Erkens: Up-to Techniques for Branching Bisimulation DESCRIPTION: Branching bisimilarity is a notion of behavioral equivalence b etween processes. To prove that two processes are branching bisimilar one s hould provide a relation (a set of pairs) containing the two processes\, su ch that the relation satisfies some properties. For recursive processes thi s relation often becomes infinite and complicated\, which may lead to awkwa rd and long ...continue reading X-ALT-DESC;FMTTYPE=text/html:Branching bisimilarity is a notion of beha vioral equivalence between processes. To prove that two processes are branc hing bisimilar one should provide a relation (a set of pairs) containing th e two processes\, such that the relation satisfies some properties. For rec ursive processes this relation often becomes infinite and complicated\, whi ch may lead to awkward and long proofs. We discuss improvements of the bran ching bisimulation proof technique that are based on the so-called 'up-to t echniques' that already exist for strong and weak bisimulation.
CATEGORIES:Colloquium LOCATION:MetaForum MF 11 GEO:51.447551;5.487453 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rick-erkens-up-to-techniq ues-for-branching-bisimulation/ END:VEVENT BEGIN:VEVENT UID:20181101T1019Z-1541067594.7602-EO-455-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20181031T161025Z LAST-MODIFIED:20181031T161522Z DTSTART;TZID=Europe/Amsterdam:20181107T140000 DTEND;TZID=Europe/Amsterdam:20181107T144500 SUMMARY: Ferry Timmers: A complete axiomatisation for probabilistic trace e quivalence DESCRIPTION: In the thesis a complete axiomatisation is given for probabili stic trace equivalence for finite\nprocesses. The axiomatisation is remarka bly complex\, which may explain that nobody formulated\nsuch a complete axi omatisation yet. X-ALT-DESC;FMTTYPE=text/html:In the thesis a sound and complete axiomat isation is given for probabilistic trace equivalence for finite probabilist ic processes. The axiomatisation is remarkably complex\, which may explain that nobody formulated such a complete axiomatisation yet.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 4 GEO:51.447551;5.487453 ORGANIZER;CN="Jan Friso Groote":MAILTO:j.f.groote@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/ferry-timmers-a-complete- axiomatisation-for-probabilistic-trace-equivalence/ END:VEVENT BEGIN:VEVENT UID:20181024T2029Z-1540412980.2235-EO-451-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20181024T060351Z LAST-MODIFIED:20181025T140007Z DTSTART;TZID=Europe/Amsterdam:20181025T133000 DTEND;TZID=Europe/Amsterdam:20181025T150000 SUMMARY: Mahmoud Talebi: Scalable performance analysis of wireless sensor n etworks X-ALT-DESC;FMTTYPE=text/html: CATEGORIES:PhD Defence LOCATION:Auditorium\, Senaatzaal GEO:51.447925;5.484193 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/mahmoud-talebi-scalable-p erformance-analysis-of-wireless-sensor-networks/ END:VEVENT BEGIN:VEVENT UID:20181017T1832Z-1539801161.1166-EO-448-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20181017T170642Z LAST-MODIFIED:20181017T170642Z DTSTART;TZID=Europe/Amsterdam:20181018T124500 DTEND;TZID=Europe/Amsterdam:20181018T131500 SUMMARY: Dragan Bosnacki: Analysis of a Boolean Model for Rheumatoid Arthri tis Using SMTSolvers DESCRIPTION: Rheumatoid Arthritis (RA) is an autoimmune disease that affect s about one percent of the world population. A medicine for RA has not yet been found\, current treatment options only decelerating the progress of th e disease. In this work a Boolean network model for RA is analysed. Two kin ds of models\, synchronous and asynchronous are compared. ...continue readi ng X-ALT-DESC;FMTTYPE=text/html:Rheumatoid Arthritis (RA) is an autoimmune
disease that affects about
one percent of the world population. A med
icine for RA has not yet been
found\, current treatment options only d
ecelerating the progress of the
disease. In this work a Boolean networ
k model for RA is analysed. Two
kinds of models\, synchronous and asyn
chronous are compared. Attractors
in the state space of the models are
identified that correspond to the
steady states and stable cycles of
the network. To this end
Satisfiability Modulo Theory (SMT) solvers an
d an adaptation of the
Tarjan’s algorithm for finding strongly connect
ed components in a graph
are used for the synchronous and asynchronous
models\, respectively. By
analysing the stability of the network nod
es potential drug targets are
identified. The results show a significa
nt overlap with similar findings
in the literature which indicates tha
t the Boolean networks can be a
feasible approach for identifying biom
arkers and drug targets.
We discuss an extension of the coordinatio
n modelling language
Paradigm. The extension is geared towards data-de
pendent interaction
among components\, where the coordination is influ
enced by possibly
distributed data. The approach is illustrated by the
well-known
example of a bakery where tickets are issued to serve clie
nts in
order. Also\, it is described how to encode Paradigm models wit
h data
in the process language of the mCRL toolset for further analysi
s
of the coordination.
Parity games are two player games with ome ga-winning conditions\, played on finite graphs. Such games play an importa nt role in verification\, satisfiability and synthesis. It is therefore imp ortant to identify algorithms that can efficiently deal with large games th at arise from such applications. In this paper\, we describe our experiment s with BDD-based implementations of four parity game solving algorithms\, v iz. Zielonka’s recursive algorithm\, the more recent Priority Promotion alg orithm\, the Fixpoint-Iteration algorithm and the automata based APT algori thm. We compare their performance on several types of random games and on a number of cases taken from the Keiren benchmark set.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tim-willemse-a-comparison -of-bdd-based-parity-game-solvers-joint-work-with-lisette-sanchez-and-wiege r-wesselink/ END:VEVENT BEGIN:VEVENT UID:20180924T1715Z-1537809323.5069-EO-438-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180924T134455Z LAST-MODIFIED:20180924T134455Z DTSTART;TZID=Europe/Amsterdam:20180927T124500 DTEND;TZID=Europe/Amsterdam:20180927T131500 SUMMARY: Thomas Neele: Fixing mistakes: narrowing the scope of quantifiers in PBESs DESCRIPTION: In the mCRL2 toolset\, formal properties are specified in the modal mu-calculus with data. This formalism allows the use of quantifiers ( ∀ and ∃) to bind data variables. However\, when a user puts a quantifier in the wrong place\, our current set of tools maybe unable to solve the resul ting model checking problem. An example ...continue reading X-ALT-DESC;FMTTYPE=text/html:In the mCRL2 toolset\, formal properties a re specified in the modal mu-calculus with data. This formalism allows the use of quantifiers (∀ and ∃) to bind data variables. However\, when a user puts a quantifier in the wrong place\, our current set of tools maybe unabl e to solve the resulting model checking problem. An example of such a poorl y written property is ∀d:D.[true*.drop(d)]false. In this talk\, I will show an automated technique to fix such errors. We perform the analysis on the level of parameterised Boolean equation systems (PBESs)\, that encode the c ombination of a model and a property.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.209 GEO:51.447551;5.487453 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/thomas-neele-fixing-mista kes-narrowing-the-scope-of-quantifiers-in-pbess/ END:VEVENT BEGIN:VEVENT UID:20180919T1138Z-1537357087.0224-EO-433-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180919T060628Z LAST-MODIFIED:20180919T060749Z DTSTART;TZID=Europe/Amsterdam:20180920T124500 DTEND;TZID=Europe/Amsterdam:20180920T131500 SUMMARY: Joel Garcia: Behavioural modelling of a wafer transport unit using mCRL2 DESCRIPTION: As a master’s student I’m doing an internship with the assigne ment to design a model for an ASML wafer transport unit. The objective is t o model the behaviour of two stages of the wafer transport unit: the wafer hanlder and the wafer stage. The model of such a machine is built by using the mCRL2 ...continue reading X-ALT-DESC;FMTTYPE=text/html:As a master’s student I'm doing an interns hip with the assignement to design a model for an ASML wafer transport unit . The objective is to model the behaviour of two stages of the wafer transp ort unit: the wafer hanlder and the wafer stage. The model of such a machin e is built by using the mCRL2 language. In this presentation\, I will expla in the progress of the work I have so far.
CATEGORIES:Colloquium LOCATION:MetaForum MF 3 GEO:51.447567;5.487443 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/joel-garcia-behavioural-m odelling-of-a-wafer-transport-unit-using-mcrl2/ END:VEVENT BEGIN:VEVENT UID:20180913T0718Z-1536823131.2466-EO-419-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180912T095905Z LAST-MODIFIED:20180912T095905Z DTSTART;TZID=Europe/Amsterdam:20180913T124500 DTEND;TZID=Europe/Amsterdam:20180913T131500 SUMMARY: Omar Alduhaiby: Learning Product Automata DESCRIPTION: Following a discussion with my colleague at University of Radb oud Joshua Moerman on learning the product of independent automata\, he pur sued the subject of independence of outputs while I the independence of inp uts. In this presentation I will summarise his recent paper titled ‘Learnin g Product Automata’ and how it relates to my work. I will ...continue readi ng X-ALT-DESC;FMTTYPE=text/html:Following a discussion with my colleague a t University of Radboud Joshua Moerman on learning the product of independe nt automata\, he pursued the subject of independence of outputs while I the independence of inputs. In this presentation I will summarise his recent p aper titled ‘Learning Product Automata’ and how it relates to my work. I wi ll present an example of a system under learning and demonstrate how its ou tput is an aggregation of a several independent outputs. The system is then separated into those smaller subsystem units\, and a modification on the l earning algorithm learns those subsystems and performs a Cartesian product on them to output the complete system’s automaton\, thus by optimizing on t he learning time. We then view the practical results from the paper and dis cuss how fruitful my approach would be.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/omar-alduhaiby-learning-p roduct-automata/ END:VEVENT BEGIN:VEVENT UID:20180905T0249Z-1536115777.6308-EO-406-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180904T204326Z LAST-MODIFIED:20180904T204326Z DTSTART;TZID=Europe/Amsterdam:20180906T124500 DTEND;TZID=Europe/Amsterdam:20180906T131500 SUMMARY: Julien Schmaltz: On the definition of block and idle for xMAS auto mata. DESCRIPTION: xMAS is a language initially proposed by Intel for architectur al modelling and verification. The main feature of xMAS is to enable a bool ean encoding of liveness that can be efficiently checked using SAT-based te chniques. xMAS is restricted to a small set of well-defined primitives. Rec ently\, Verbeek et al. extended this approach to state machines. In ...cont inue reading X-ALT-DESC;FMTTYPE=text/html:xMAS is a language initially proposed by I ntel for architectural modelling and verification. The main feature of xMAS is to enable a boolean encoding of liveness that can be efficiently checke d using SAT-based techniques. xMAS is restricted to a small set of well-def ined primitives. Recently\, Verbeek et al. extended this approach to state machines. In this talk\, I will recall the definitions they propose for the main boolean variables (block and idle) of the encoding. I will then show two counter-examples to these definition. Together with Alexander\, we are currently working on fixing these definitions. I will discuss our current h int. This will be quite sketchy as we do not really have a satisfactory sol ution yet.
CATEGORIES:Colloquium LOCATION:MetaForum MF 12 GEO:51.447551;5.487453 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/julien-schmaltz-on-the-de finition-of-block-and-idle-for-xmas-automata/ END:VEVENT BEGIN:VEVENT UID:20180808T2033Z-1533760408.1177-EO-389-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180808T130322Z LAST-MODIFIED:20180813T135934Z DTSTART;TZID=Europe/Amsterdam:20180814T160000 DTEND;TZID=Europe/Amsterdam:20180814T164500 SUMMARY: Lisette Sanchez: Learning software behavior through active automat a learning with data DESCRIPTION: The topic of the thesis is find out whether the LS* learning a lgorithm\, that can learn register automata with abstract data parameters f rom actual software is practically applicable in an industrial context. The algorithm can for instance learn a queue with limited size that stores arb itrary natural numbers. The conclusion is that indeed practical software .. .continue reading X-ALT-DESC;FMTTYPE=text/html:The topic of the thesis is find out whethe r the LS* learning algorithm\, that can learn register automata with abstra ct data parameters from actual software is practically applicable in an ind ustrial context. The algorithm can for instance learn a queue with limited size that stores arbitrary natural numbers.
The conclusion is that i ndeed practical software can be learned\, outperforming the learning of aut omata with concrete data. Scale remains a problem\, though.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 13 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/lisette-sanchez-learning- software-behavior-through-active-automata-learning-with-data/ END:VEVENT BEGIN:VEVENT UID:20180718T0944Z-1531907068.7232-EO-377-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180718T091528Z LAST-MODIFIED:20180718T091528Z DTSTART;TZID=Europe/Amsterdam:20180724T100000 DTEND;TZID=Europe/Amsterdam:20180724T104500 SUMMARY: Lois Nijland: Adding sequential composition and termination to the linear time – branching time spectrum DESCRIPTION: Van Glabbeek presented the linear time – branching time spectr um of behavioural semantics and gave sound\, ground-complete axiomatisation s for the process theory BCCSP. Groote and Chen et al. proved for the seman tics in the spectrum whether there exist finite axiomatisations that are om ega-complete. We add termination and sequential composition to the spectrum by studying the ...continue reading X-ALT-DESC;FMTTYPE=text/html:Van Glabbeek presented the linear time - b ranching time spectrum of behavioural semantics and gave sound\, ground-com plete axiomatisations for the process theory BCCSP. Groote and Chen et al. proved for the semantics in the spectrum whether there exist finite axiomat isations that are omega-complete. We add termination and sequential composi tion to the spectrum by studying the semantics for the process theories BSP and TSP. We provide a template for proving soundness and ground-completene ss for BSP and apply this template to BSP modulo trace semantics. We prove that for BSP modulo ready simulation semantics with a finite number of acti ons of at least one a finite basis does not exist by giving an infinite fam ily of equations which are sound but cannot be derived from a finite collec tion of axioms that are sound for BSP modulo ready simulation semantics. We show that we cannot use this family of equations to prove whether TSP modu lo ready simulation semantics affords a finite basis. Finally\, we prove th at TSP modulo bisimulation is omega-complete when there is at least one act ion.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 14 GEO:51.447532;5.487437 ORGANIZER;CN="Bas Luttik":MAILTO:s.p.luttik@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/lois-nijland-adding-seque ntial-composition-and-termination-to-the-linear-time-branching-time-spectru m/ END:VEVENT BEGIN:VEVENT UID:20180717T1144Z-1531827868.9545-EO-376-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180717T095712Z LAST-MODIFIED:20180717T095712Z DTSTART;TZID=Europe/Amsterdam:20180723T150000 DTEND;TZID=Europe/Amsterdam:20180723T154500 SUMMARY: Astrid Belder: Decidability of bisimilarity and axiomatisation for sequential processes in the presence of intermediate termination DESCRIPTION: An alternative semantics for sequential composition in a setti ng with intermediate termination was proposed in a recent article by Baeten \, Luttik and Yang. We consider two open questions regarding sequential pro cesses with intermediate termination that use the revised semantics for seq uential composition (TSP\;). First\, a ground-complete axiomatisation is pr oposed for TSP\;\, extended with an auxiliary ...continue reading X-ALT-DESC;FMTTYPE=text/html:An alternative semantics for sequential co mposition in a setting with intermediate termination was proposed in a rece nt article by Baeten\, Luttik and Yang. We consider two open questions rega rding sequential processes with intermediate termination that use the revis ed semantics for sequential composition (TSP\;). First\, a groun d-complete axiomatisation is proposed for TSP\;\, extended with an auxiliary operator that is used to remove intermediate termination from terms. Additionally\, it is shown that TSP\; does not afford a g round-complete axiomatisation with respect to bisimilarity without an auxil iary operator. Second\, we prove that bisimilarity is decidable for process es definable by finite guarded recursive specification over TSP\;. This is done by showing that every guarded recursive TSP\;-sp ecification can be transformed to a normal form\, which allows us to elimin ate redundant intermediate termination from processes. Using this normal fo rm\, the existing decidability proofs for context-free processes without in termediate termination\, can be adapted for TSP\;.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 14 GEO:51.447532;5.487437 ORGANIZER;CN="Bas Luttik":MAILTO:s.p.luttik@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/astrid-belder-decidabilit y-of-bisimilarity-and-axiomatisation-for-sequential-processes-in-the-presen ce-of-intermediate-termination/ END:VEVENT BEGIN:VEVENT UID:20180704T0251Z-1530672715.5931-EO-367-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180703T125022Z LAST-MODIFIED:20180703T125022Z DTSTART;TZID=Europe/Amsterdam:20180705T124500 DTEND;TZID=Europe/Amsterdam:20180705T133000 SUMMARY: Tatiana Shmeleva: Modeling Networks with Infinite Petri Nets DESCRIPTION: A composition and analysis technique is developed for investig ation of infinite Petri nets with regular structure\, introduced for modeli ng networks\, clusters and computing grids\, that also concerns cellular au tomata and biological systems. A case study of a square grid structure comp osition and analysis is presented. Parametric specification of Petri nets\, parametric representation of infinite systems ...continue reading X-ALT-DESC;FMTTYPE=text/html:A composition and analysis technique is de veloped for investigation of infinite Petri nets with regular structure\, i ntroduced for modeling networks\, clusters and computing grids\, that also concerns cellular automata and biological systems. A case study of a square grid structure composition and analysis is presented. Parametric specifica tion of Petri nets\, parametric representation of infinite systems for the calculation of place/transition invariants\, and solving them in parametric form allowed the invariance proof for infinite Petri net models. Some addi tional analysis techniques based on graphs of transmissions and blockings a re presented. Further generalization on multidimensional structures such as hypercube and hypertorus have been implemented. Generators of Petri net mo dels have been developed and put on GitHub for public use. Complex deadlock s are disclosed and a possibility of network blocking via ill-intentioned t raffic revealed. Quality of service in modern networks and numerical parame ters of blocking networks with disguised ill-intentioned traffic are invest igated using reenterable models in the form of colored Petri nets.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tatiana-shmeleva-modeling -networks-with-infinite-petri-nets/ END:VEVENT BEGIN:VEVENT UID:20180702T1919Z-1530559177.5363-EO-366-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180702T173658Z LAST-MODIFIED:20180708T174452Z DTSTART;TZID=Europe/Amsterdam:20180702T100000 DTEND;TZID=Europe/Amsterdam:20180702T104500 SUMMARY: Marijn Rol: Verification of ASD multi-component systems in mCRL2 DESCRIPTION: Analytical Software Design (ASD) assists the creation of softw are systems. Systems designed in ASD are composed of multiple components in order to divide the complexity of the whole system over them. The verifica tion of system properties and requirements is limited to the scope of singl e components\, disallowing the verification of end-to-end properties. We pr esent an ...continue reading X-ALT-DESC;FMTTYPE=text/html:Analytical Software Design (ASD) assists t he creation of software systems. Systems designed in ASD are composed of mu ltiple components in order to divide the complexity of the whole system ove r them. The verification of system properties and requirements is limited t o the scope of single components\, disallowing the verification of end-to-e nd properties. We present an approach for the verification of end-to-end pr operties on multi-component systems. This provides a higher confidence in t he functional correctness and reliability. A system based on a real-life AS D model serves as use-case for the proposed approach. Results show that ver ification of multi-component systems can be done through mCRL2\, but scalab ility issues are observed as larger systems are verified.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 13 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/marijn-rol-verification-o f-asd-multi-component-systems-in-mcrl2/ END:VEVENT BEGIN:VEVENT UID:20180626T1000Z-1530007224.222-EO-364-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180626T092822Z LAST-MODIFIED:20180626T092822Z DTSTART;TZID=Europe/Amsterdam:20180628T124500 DTEND;TZID=Europe/Amsterdam:20180628T133000 SUMMARY: Maurice Laveaux: Verification in mCRL2 using multi-threaded algori thms: The term library. DESCRIPTION: Currently the mCRL2 toolset uses sequential algorithms exclusi vely. From these sequential algorithms the term rewriting algorithm often h as the most prominent run-time cost. Developing a parallel term rewriting a lgorithm requires an efficient term library. This library facilitates the c reation and destruction of terms. In this presentation I will present sever al challenges that I encountered while ...continue reading X-ALT-DESC;FMTTYPE=text/html:Currently the mCRL2 toolset uses sequentia l algorithms exclusively. From these sequential algorithms the term rewriti ng algorithm often has the most prominent run-time cost. Developing a paral lel term rewriting algorithm requires an efficient term library. This libra ry facilitates the creation and destruction of terms. In this presentation I will present several challenges that I encountered while trying to develo p an efficient term library. Followed by several observations that lead to an envisioned design for a parallel term rewriting algorithm.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/maurice-laveaux-verificat ion-in-mcrl2-using-multi-threaded-algorithms-the-term-library/ END:VEVENT BEGIN:VEVENT UID:20180618T2021Z-1529353314.2045-EO-361-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180610T174747Z LAST-MODIFIED:20180708T174543Z DTSTART;TZID=Europe/Amsterdam:20180611T110000 DTEND;TZID=Europe/Amsterdam:20180611T123000 SUMMARY: Fei Yang: A Theory of Executability —- With a Focus on the Express ivity of Process Calculi X-ALT-DESC;FMTTYPE=text/html: CATEGORIES:PhD Defence LOCATION:Auditorium\, Senaatzaal GEO:51.447925;5.484193 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/fei-yang-a-theory-of-exec utability-with-a-focus-on-the-expressivity-of-process-calculi/ END:VEVENT BEGIN:VEVENT UID:20180604T1504Z-1528124668.0894-EO-349-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180604T101812Z LAST-MODIFIED:20180604T101812Z DTSTART;TZID=Europe/Amsterdam:20180607T124500 DTEND;TZID=Europe/Amsterdam:20180607T133000 SUMMARY: Hans Zantema: Making a MOOC DESCRIPTION: To make my lectures Automated Reasoning more accessible worldw ide\, I started developing a series of MOOCs (massive open online course). Professional facilities are provided by EIT Digital and Coursera. At the mo ment 21 lectures have been recorded\, each covering around 10 minutes\, by which the first MOOC on satisfiability is nearly finished. Roughly this MOO C ...continue reading X-ALT-DESC;FMTTYPE=text/html:To make my lectures Automated Reasoning mo re accessible worldwide\, I started developing a series of MOOCs (massive o pen online course). Professional facilities are provided by EIT Digital and Coursera. At the moment 21 lectures have been recorded\, each covering aro und 10 minutes\, by which the first MOOC on satisfiability is nearly finish ed. Roughly this MOOC covers the same material on SAT/SMT as in my course A utomated Reasoning\, being around 40% of the course. In the talk this exper ience will be discussed\, in particular comparing this format with usual te aching. Also one or two of the lectures will be shown.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Alexander Fedotov":MAILTO:a.fedotov@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/hans-zantema-making-a-moo c/ END:VEVENT BEGIN:VEVENT UID:20180530T2325Z-1527722717.5901-EO-343-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180530T095339Z LAST-MODIFIED:20180530T095339Z DTSTART;TZID=Europe/Amsterdam:20180531T124500 DTEND;TZID=Europe/Amsterdam:20180531T133000 SUMMARY: Mahmoud Talebi: First-order Moment Closure Approximations for Midd le-Sized Systems with Non-linear Rates DESCRIPTION: In this presentation I talk about the problem of approximating the behaviour of middle-sized population models involving non-linear rates . I describe a number of systems\, each with a very different non-linear be haviour\, and then show that the binomial and Poisson moment closure approx imations have the potential to accurately represent the expected behaviour of these models. ...continue reading X-ALT-DESC;FMTTYPE=text/html:In this presentation I talk about the prob lem of approximating the behaviour of middle-sized population models involv ing non-linear rates. I describe a number of systems\, each with a very dif ferent non-linear behaviour\, and then show that the binomial and Poisson m oment closure approximations have the potential to accurately represent the expected behaviour of these models. I then compare the two approximation m ethods to the mean field and normal moment closure approximations in terms of applicability and accuracy\, in addition to investigating their empirica l distributions\, to derive conclusions regarding their strengths and drawb acks.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Fei Yang":MAILTO:f.yang@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/mahmoud-talebi-first-orde r-moment-closure-approximations-for-middle-sized-systems-with-non-linear-ra tes/ END:VEVENT BEGIN:VEVENT UID:20180524T0318Z-1527131920.8652-EO-339-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180523T140755Z LAST-MODIFIED:20180523T140755Z DTSTART;TZID=Europe/Amsterdam:20180524T124500 DTEND;TZID=Europe/Amsterdam:20180524T133000 SUMMARY: Jan Friso Groote: An attempt to reap the benefits of processor ari thmetic without loosing the advantages of the numbers represented as rewrit e systems DESCRIPTION: In mCRL2 numbers are represented using rewrite systems. The re presentation is based on binary operations. This is relatively concise but far less efficient than 64 bits operations on numbers available in processo rs. I will explain ongoing work towards an attempt to reap the benefits of processor arithmetic without loosing the advantages of the numbers represen ted ...continue reading X-ALT-DESC;FMTTYPE=text/html:In mCRL2 numbers are represented using rew rite systems. The representation is based on binary operations. This is rel atively concise but far less efficient than 64 bits operations on numbers a vailable in processors. I will explain ongoing work towards an attempt to r eap the benefits of processor arithmetic without loosing the advantages of the numbers represented as rewrite systems.
CATEGORIES:Colloquium LOCATION:MetaForum MF 3 GEO:51.447567;5.487443 ORGANIZER;CN="Fei Yang":MAILTO:f.yang@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-friso-groote-an-attem pt-to-reap-the-benefits-of-processor-arithmetic-without-loosing-the-advanta ges-of-the-numbers-represented-as-rewrite-systems/ END:VEVENT BEGIN:VEVENT UID:20180516T0043Z-1526431424.649-EO-321-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180515T105157Z LAST-MODIFIED:20180516T110034Z DTSTART;TZID=Europe/Amsterdam:20180518T100000 DTEND;TZID=Europe/Amsterdam:20180518T104500 SUMMARY: Ziad Ben Snaiba: Logics and Algorithms for Product and Family-Base d Analysis of Software Product Lines DESCRIPTION: Most model checking techniques consider single systems. With t he rise of using Software Product Line Engineering (SPLE) for critical syst ems there is a need for model checking techniques applicable to Software Pr oduct Lines (SPL). Inspired by Classen et al. we try to devise logics to fo rmally describe properties suitable for the analysis of SPLs. We ...continu e reading X-ALT-DESC;FMTTYPE=text/html:Most model checking techniques consider si ngle systems. With the rise of using Software Product Line Engineering (SPL E) for critical systems there is a need for model checking techniques appli cable to Software Product Lines (SPL). Inspired by Classen et al. we try to devise logics to formally describe properties suitable for the analysis of SPLs. We attempt to provide logics for both product-based and family-based model checking based on CTL. These logics allow both nested (nCTL) and sin gle (sCTL) product-family restrictions on the specification of the SPL. Als o\, we provide an equivalence between nCTL and sCTL\, and algorithms for bo th product-based and family-based model checking using sCTL and nCTL. A sma ll toolset to use the algorithms in practice will be provided.
CATEGORIES:MSc Defence LOCATION:Flux 1.10 GEO:51.447291;5.492092 ORGANIZER;CN="Erik de Vink":MAILTO:e.p.d.vink@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/ziad-ben-snaiba-logics-an d-algorithms-for-product-and-family-based-analysis-of-software-product-line s/ END:VEVENT BEGIN:VEVENT UID:20180517T0431Z-1526531508.0978-EO-330-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180516T145617Z LAST-MODIFIED:20180516T145617Z DTSTART;TZID=Europe/Amsterdam:20180517T124500 DTEND;TZID=Europe/Amsterdam:20180517T133000 SUMMARY: Bas Luttik: Modelling and Analysing ERTMS Hybrid Level 3 with the mCRL2 toolset DESCRIPTION: ERTMS Hybrid Level 3 is a recent proposal for a train control system specification that serves to increase the capacity of the railway ne twork by allowing multiple trains on a single track section. We have formal ly modelled and analysed the principles of ERTMS Hybrid Level 3 in mCRL2. O ur analysis has resulted in suggestions for ...continue reading X-ALT-DESC;FMTTYPE=text/html:ERTMS Hybrid Level 3 is a recent proposal for a train control system specification that serves to increase the capaci ty of the railway network by allowing multiple trains on a single track sec tion.
We have formally modelled and analysed the principles of ERTMS Hybrid Level 3 in mCRL2.
Our analysis has resulted in suggestions for improvement of the principles that will be taken into account in the next version of the specification.
In this talk I’ll introduce ERTMS Hybri d Level 3\, I’ll discuss how we have obtained an mCRL2 model (or actually s everal mCRL2 models) for it\, and I’ll report on and explain some of the re sults of our analysis.
(Based on joint work with Maarten Bartholomeus and Tim Willemse.)
CATEGORIES:Colloquium LOCATION:MetaForum 3.141 GEO:51.447567;5.487443 ORGANIZER;CN="Fei Yang":MAILTO:f.yang@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/bas-luttik-modelling-and- analysing-ertms-hybrid-level-3-with-the-mcrl2-toolset/ END:VEVENT BEGIN:VEVENT UID:20180502T2047Z-1525294050.3391-EO-314-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180502T145808Z LAST-MODIFIED:20180502T145808Z DTSTART;TZID=Europe/Amsterdam:20180503T124500 DTEND;TZID=Europe/Amsterdam:20180503T133000 SUMMARY: Muhammad Osama: SIGmA: SAT Simplifications on GPU Architecture DESCRIPTION: The growing scale of industrial applications encoded to Boolea n Satisfiability (SAT) problems imposed the researchers to practice SAT sim plification as an imperative requirement for any SAT solver. In this talk\, I will discuss how GPU can be utilized to perform variable and subsumption eliminations in parallel. Benchmarks show that our proposed simplifier (SI GmA) achieved an ...continue reading X-ALT-DESC;FMTTYPE=text/html:The growing scale of industrial applicatio ns encoded to Boolean Satisfiability (SAT) problems imposed the researchers to practice SAT simplification as an imperative requirement for any SAT so lver. In this talk\, I will discuss how GPU can be utilized to perform vari able and subsumption eliminations in parallel. Benchmarks show that our pro posed simplifier (SIGmA) achieved an acceleration of 250x over SatELite. Re garding SAT solving\, SIGmA outperformed SatELite in terms of problems solv ed faster when combined with MiniSat by a factor of 77%. Moreover\, MiniSat was more effective than Lingeling by 73.3% in solving simplified formulas given by SIGmA.
CATEGORIES:Colloquium LOCATION:MetaForum MF 5.117 GEO:51.447567;5.487443 ORGANIZER;CN="Fei Yang":MAILTO:f.yang@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/muhammad-osama-sigma-sat- simplifications-on-gpu-architecture/ END:VEVENT BEGIN:VEVENT UID:20180425T2141Z-1524692510.7463-EO-305-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180425T164504Z LAST-MODIFIED:20180425T164504Z DTSTART;TZID=Europe/Amsterdam:20180426T124500 DTEND;TZID=Europe/Amsterdam:20180426T133000 SUMMARY: Erik de Vink: On the BEM algorithm for deciding strong probabilist ic bisimulation DESCRIPTION: An algorithm for deciding strong probabilistic bisimulation\, referred to as BEM\, was proposed in 2000 by Baier\, Engelen\, and Majster- Cederbaum claiming O(m n log n) complexity for the construction of the bisi mulation equivalence classes of a probabilistic automaton. Besides so-calle d Ordered Balanced Trees\, the BEM algorithm uses straightforward datastruc tures. More recently\, Groote and Rivera Verduzco ...continue reading X-ALT-DESC;FMTTYPE=text/html:An algorithm for deciding strong probabili stic bisimulation\, referred to as BEM\, was proposed in 2000 by Baier\, En gelen\, and Majster-Cederbaum claiming O(m n log n) complexity for the cons truction of the bisimulation equivalence classes of a probabilistic automat on. Besides so-called Ordered Balanced Trees\, the BEM algorithm uses strai ghtforward datastructures. More recently\, Groote and Rivera Verduzco propo sed an algorithm\, referred to as GRV\, of the same complexity based on mor e intricate data structures. The running times of GRV\, however\, are aroun d 1000 better than those of BEM\, on a standard set of benchmarks. Today we take a closer look at how the BEM algorithm works\, seeking explanation fo r the difference in performance compared to the GRV algorithm\, leaving GRV for another occasion.
CATEGORIES:Colloquium LOCATION:MetaForum MF 5.117 GEO:51.447567;5.487443 ORGANIZER;CN="Fei Yang":MAILTO:f.yang@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/erik-de-vink-on-the-bem-a lgorithm-for-deciding-strong-probabilistic-bisimulation/ END:VEVENT BEGIN:VEVENT UID:20180411T1706Z-1523466407.8255-EO-301-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180411T101312Z LAST-MODIFIED:20180411T101312Z DTSTART;TZID=Europe/Amsterdam:20180412T124500 DTEND;TZID=Europe/Amsterdam:20180412T133000 SUMMARY: Wieger Wesselink: Counterexamples in the mCRL2 toolset DESCRIPTION: A long-standing problem in the mCRL2 toolset was the poor supp ort for generating counterexamples of model checking properties. Over the c ourse of the last year an effort has been made to finally tackle this probl em. It is now possible to generate counterexamples and translate them back to the original model. The first applications of this ...continue reading X-ALT-DESC;FMTTYPE=text/html:A long-standing problem in the mCRL2 tools et was the poor support for generating counterexamples of model checking pr operties. Over the course of the last year an effort has been made to final ly tackle this problem. It is now possible to generate counterexamples and translate them back to the original model. The first applications of this n ew technique look very promising. In this talk I will roughly explain how t he generation of counterexamples works\, and give some examples. I will als o explain my thoughts about how to implement such an algorithm.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Fei Yang":MAILTO:f.yang@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/wieger-wesselink-countere xamples-in-the-mcrl2-toolset/ END:VEVENT BEGIN:VEVENT UID:20180404T0856Z-1522832197.7382-EO-297-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180404T083800Z LAST-MODIFIED:20180404T083800Z DTSTART;TZID=Europe/Amsterdam:20180405T124500 DTEND;TZID=Europe/Amsterdam:20180405T133000 SUMMARY: Rodin Aarsen: Static Analysis on Legacy Software DESCRIPTION: Legacy software is a prominent bottleneck in modern industry: it is hard and costly to maintain\, yet contains valuable knowledge not ava ilable elsewhere. Static analysis provides insight by extracting facts from existing code bases. In this talk\, I will address complications and progr ess on my static analysis tooling in Rascal. X-ALT-DESC;FMTTYPE=text/html:Legacy software is a prominent bottleneck in modern industry: it is hard and costly to maintain\, yet contains valuab le knowledge not available elsewhere. Static analysis provides insight by e xtracting facts from existing code bases. In this talk\, I will address com plications and progress on my static analysis tooling in Rascal.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Mahmoud Talebi":MAILTO:m.talebi@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rodin-aarsen-static-analy sis-on-legacy-software/ END:VEVENT BEGIN:VEVENT UID:20180328T1111Z-1522235509.5536-EO-285-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180328T075258Z LAST-MODIFIED:20180328T075258Z DTSTART;TZID=Europe/Amsterdam:20180329T124500 DTEND;TZID=Europe/Amsterdam:20180329T133000 SUMMARY: Pieter Hijma: Programming many-cores with the “Stepwise-refinement for performance” methodology DESCRIPTION: The main goal of this talk is to introduce my research to the FSA group and discuss possibilities and opportunities on how this research can be combined with model checking in general and with mCRL2 in particular in the context of the recently accepted TOP project AVVA (Accelerated Veri fication and Verified Acceleration). I will present ...continue reading X-ALT-DESC;FMTTYPE=text/html:The main goal of this talk is to introduce my research to the FSA group and discuss possibilities and opportunities o n how this research can be combined with model checking in general and with mCRL2 in particular in the context of the recently accepted TOP project AV VA (Accelerated Verification and Verified Acceleration). I will present my programming system Many-Core Levels\, the principles behind it\, and how it exposes a methodology to programmers that we call "Stepwise-refinement for performance". Many-core hardware is targeted specifically at obtaining hig h performance. However\, obtaining high performance is challenging because hardware-specific details have to be taken into account. In our system\, pr ogrammers can define and choose their own level of abstraction: higher leve ls for readability and portability\, and user-defined lower levels to incor portate more hardware-specific details to obtain higher performance.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Fei Yang":MAILTO:f.yang@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/pieter-hijma-programming- many-cores-with-the-stepwise-refinement-for-performance-methodology/ END:VEVENT BEGIN:VEVENT UID:20180323T0135Z-1521768934.5068-EO-277-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180321T155410Z LAST-MODIFIED:20180321T155410Z DTSTART;TZID=Europe/Amsterdam:20180322T124500 DTEND;TZID=Europe/Amsterdam:20180322T133000 SUMMARY: Fei Yang: Two Problems on Context-free Processes and Pushdown Proc esses DESCRIPTION: In this talk\, I will discuss two problems on examining in whi ch cases context-free processes and pushdown processes are the same. In par ticular\, we depart from the well-known case of language equivalence and in stead look at processes using process theory and more fine-grained equivale nces\, such as branching bisimulation and contrasimulation. We identify two difficulties when ...continue reading X-ALT-DESC;FMTTYPE=text/html:In this talk\, I will discuss two problems on examining in which cases context-free processes and pushdown processes are the same. In particular\, we depart from the well-known case of languag e equivalence and instead look at processes using process theory and more f ine-grained equivalences\, such as branching bisimulation and contrasimulat ion. We identify two difficulties when looking at process specifications: h ead-recursion and transparency. Here two new results are achieved: we prove that when excluding transparency\, context-free processes and pushdown pro cesses are equivalent up to at least divergence-insensitive variant of bran ching bisimilarity. When including transparency\, we prove that they are eq uivalent up to at least contrasimulation.
This is a joint work with J os Baeten and Zeno de Hoop.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Fei Yang":MAILTO:f.yang@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/fei-yang-two-problems-on- context-free-processes-and-pushdown-processes/ END:VEVENT BEGIN:VEVENT UID:20180314T2117Z-1521062277.7493-EO-272-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180314T170321Z LAST-MODIFIED:20180314T170321Z DTSTART;TZID=Europe/Amsterdam:20180315T124500 DTEND;TZID=Europe/Amsterdam:20180315T133000 SUMMARY: Omar Alzuhaibi: Lessons Learned from Model Learning of Legacy Soft ware DESCRIPTION: Legacy software is one of the most common struggles of the cur rent software industry\, being costly to maintain yet essential for the ong oing industrial process. This makes refactoring especially risky. The black -box reverse engineering technique we call model learning comes to aid. As pleasant in theory\, as perilous in application it is however. This present ation ...continue reading X-ALT-DESC;FMTTYPE=text/html:Legacy software is one of the most common struggles of the current software industry\, being costly to maintain yet e ssential for the ongoing industrial process. This makes refactoring especia lly risky. The black-box reverse engineering technique we call model learni ng comes to aid. As pleasant in theory\, as perilous in application it is h owever. This presentation walks through some of these dangers and demonstra tes lessons I learned from analysing certain legacy components at Philips H ealthcare.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Fei Yang":MAILTO:f.yang@tue.nl URL;VALUE=URI:https://fsa.win.tue.nl/events/event/omar-alzuhaibi-lessons-le arned-from-model-learning-of-legacy-software/ END:VEVENT BEGIN:VEVENT UID:20180312T1759Z-1520877567.5233-EO-266-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180312T174345Z LAST-MODIFIED:20180708T174543Z DTSTART;TZID=Europe/Amsterdam:20180313T100000 DTEND;TZID=Europe/Amsterdam:20180313T104500 SUMMARY: Ruud van Vijfeijken: Performance evaluation of network on chip usi ng formal models DESCRIPTION: Because of the increase in complexity of System-on-Chip (SoC) networks by using Network-on-Chip (NoC)\, latency has become a major issue in the design and validation. There are formal and mathematical methods to determine latency bounds of a NoC\, such as network calculus\, and low leve l simulation environments\, such as a cycle-accurate simulation of the Regi ster ...continue reading X-ALT-DESC;FMTTYPE=text/html:Because of the increase in complexity of S ystem-on-Chip (SoC) networks by using Network-on-Chip (NoC)\, latency has b ecome a major issue in the design and validation. There are formal and math ematical methods to determine latency bounds of a NoC\, such as network cal culus\, and low level simulation environments\, such as a cycle-accurate si mulation of the Register Transfer Level. This thesis proposes a method that abstracts away from low level analysis and uses a formal model to analyze the latency bounds of a given NoC. By applying different traffic schemes we can determine worst case latency of some case studies\, which include a 8- node Spidergon with cache coherency and a TornadoNoC architecture.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 3 GEO:51.447567;5.487443 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/ruud-van-vijfeijken-perfo rmance-evaluation-of-network-on-chip-using-formal-models/ END:VEVENT BEGIN:VEVENT UID:20180308T1147Z-1520509676.6478-EO-260-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180308T105137Z LAST-MODIFIED:20180708T174625Z DTSTART;TZID=Europe/Amsterdam:20180308T124500 DTEND;TZID=Europe/Amsterdam:20180308T133000 SUMMARY: Alexander Fedotov: Automatic Generation of Hardware Checkers from Formal Micro-architectural Specifications DESCRIPTION: To manage design complexity\, high-level models are used to ev aluate the functionality and performance of design solutions. There is a si gnificant gap between these high-level models and the Register Transfer Lev el (RTL) implementations actually produced by designers. We address the cha llenge of bridging this gap\, namely\, relating abstract specifications to RTL implementations. An important feature ...continue reading X-ALT-DESC;FMTTYPE=text/html:To manage design complexity\, high-level m odels are used to evaluate the functionality and performance of design solu tions. There is a significant gap between these high-level models and the R egister Transfer Level (RTL) implementations actually produced by designers . We address the challenge of bridging this gap\, namely\, relating abstrac t specifications to RTL implementations. An important feature of our propos ed approach is to support non-deterministic specifications. From such a non -deterministic model\, we automatically compute a representation of its obs ervable behaviour. We then turn this representation into a System Verilog c hecker. The checker is connected to the input and output interfaces of the RTL implementation. The resulting combination is given to a commercial EDA tool to prove that the specification simulates the implementation. Our meth od is implemented for the formal micro-architectural description language ( MaDL) ¨C an extension of the xMAS formalism originally proposed by Intel ¨C and exemplified on several examples.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/alexander-fedotov-automat ic-generation-of-hardware-checkers-from-formal-micro-architectural-specific ations/ END:VEVENT BEGIN:VEVENT UID:20180227T1505Z-1519743947.596-EO-244-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180227T123012Z LAST-MODIFIED:20180708T174625Z DTSTART;TZID=Europe/Amsterdam:20180301T124500 DTEND;TZID=Europe/Amsterdam:20180301T133000 SUMMARY: Thomas Neele: Proof searching in infinite parity games DESCRIPTION: One way to perform model checking of a model mu-calculus formu la on a linear process is to transform the problem into a parity game. The solution to this parity game also provides an answer to the original model checking question. However\, for infinite systems\, such as real-time syste ms\, the parity game corresponding to most non-trivial ...continue reading X-ALT-DESC;FMTTYPE=text/html:One way to perform model checking of a mod el mu-calculus formula on a linear process is to transform the problem into a parity game. The solution to this parity game also provides an answer to the original model checking question. However\, for infinite systems\, suc h as real-time systems\, the parity game corresponding to most non-trivial properties is also of infinite size. To still reason about its solution\, w e developed a technique that can efficiently search for a witness or counte r-example in infinite parity games. The search is based on information cont ained in a given parameterised Boolean equation system (PBES) that encodes our model checking problem.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/thomas-neele-proof-search ing-in-infinite-parity-games/ END:VEVENT BEGIN:VEVENT UID:20180208T0057Z-1518051434.7578-EO-228-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180207T201121Z LAST-MODIFIED:20180708T174625Z DTSTART;TZID=Europe/Amsterdam:20180208T124500 DTEND;TZID=Europe/Amsterdam:20180208T133000 SUMMARY: Julien Schmaltz: Formal analysis using the MaDL Whiteboard — a dem o DESCRIPTION: I will give a demo of the tool we are developing based on our research about formal analysis of micro-architectures. Going through exampl es I will illustrate the main modelling and verification features we curren tly have implemented. An interesting aspect is that we can model asynchrono us systems. This part still requires some research and might generate ...co ntinue reading X-ALT-DESC;FMTTYPE=text/html:I will give a demo of the tool we are deve loping based on our research about formal analysis of micro-architectures. Going through examples I will illustrate the main modelling and verificatio n features we currently have implemented. An interesting aspect is that we can model asynchronous systems. This part still requires some research and might generate some discussion :-)
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/julien-schmaltz-formal-an alysis-using-the-madl-whiteboard-a-demo/ END:VEVENT BEGIN:VEVENT UID:20180201T0356Z-1517457408.1893-EO-227-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180131T110544Z LAST-MODIFIED:20180708T174625Z DTSTART;TZID=Europe/Amsterdam:20180201T124500 DTEND;TZID=Europe/Amsterdam:20180201T133000 SUMMARY: Hans Zantema: Some basics of SAT/SMT solving DESCRIPTION: SAT/SMT solving is a general technique for automatically findi ng solutions for a wide range of problems. In this talk some basics of the underlying techniques are presented. First we focus on SAT = satisfiability of pure propositional formulas. Next we see how the same techniques can be extended to SMT = satisfiability modulo theories\, in ...continue reading X-ALT-DESC;FMTTYPE=text/html:SAT/SMT solving is a general technique for automatically finding solutions for a wide range of problems. In this talk some basics of the underlying techniques are presented. First we focus on SAT = satisfiability of pure propositional formulas. Next we see how the sa me techniques can be extended to SMT = satisfiability modulo theories\, in particular for the theory of linear inequalities. That is\, the formulas ar e not only composed from Boolean variables and propositional operations\, b ut also from linear inequalities over numbers.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/hans-zantema-some-basics- of-sat-smt-solving/ END:VEVENT BEGIN:VEVENT UID:20180124T0632Z-1516775549.9902-EO-225-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180123T171749Z LAST-MODIFIED:20180708T174625Z DTSTART;TZID=Europe/Amsterdam:20180130T140000 DTEND;TZID=Europe/Amsterdam:20180130T144500 SUMMARY: Maurice Laveaux: Abstracting real-valued parameters in parameteris ed Boolean equation systems DESCRIPTION: The mCRL2 tool-set utilizes parameterised boolean equation sys tems to verify formulas from modal mu-calculus on models written in the min imal common representation language (mCRL2). For models of real-timed syste ms this introduces real-valued parameters in these equation systems. Solvin g parameterised boolean equation systems with real-valued parameters is not possible in most cases. We will show that ...continue reading X-ALT-DESC;FMTTYPE=text/html:The mCRL2 tool-set utilizes parameterised boolean equation systems to verify formulas from modal mu-calculus on model s written in the minimal common representation language (mCRL2). For models of real-timed systems this introduces real-valued parameters in these equa tion systems. Solving parameterised boolean equation systems with real-valu ed parameters is not possible in most cases.
We will show that a regi on abstraction\, derived from a similar notion defined for timed automata\, can also be applied to parameterised boolean equation systems. An alternat ive abstraction where regions are combined into so-called zones will be def ined as well. In some cases this approach works better in practice. However \, it is also more restricted in the type of model checking questions that it can answer. Finally\, we will define suitable representations for region s and zones that have also been implemented using the mCRL2 language.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 14 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/maurice-laveaux-abstracti ng-real-valued-parameters-in-parameterised-boolean-equation-systems/ END:VEVENT BEGIN:VEVENT UID:20180119T1918Z-1516389498.7558-EO-216-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180119T082931Z LAST-MODIFIED:20181031T192554Z DTSTART;TZID=Europe/Amsterdam:20180125T140000 DTEND;TZID=Europe/Amsterdam:20180125T144500 SUMMARY: Roxana Paval: Modeling and Verifying Concurrent Data Structures DESCRIPTION: Concurrent data structures can be used to communicate between parallel processes in a system. The challenge in manipulating these objects arises from the many possible ways in which the processes can interleave. To ensure correct executions\, the system should fulfill linearizability. V erifying linearizability consists of checking that every concurrent executi on is equivalent to some sequential ...continue reading X-ALT-DESC;FMTTYPE=text/html:Concurrent data structures can be used to communicate between parallel processes in a system. The challenge in manipu lating these objects arises from the many possible ways in which the proces ses can interleave. To ensure correct executions\, the system should fulfil l linearizability. Verifying linearizability consists of checking that ever y concurrent execution is equivalent to some sequential execution that resp ects the runtime ordering of methods. This work proposes building two proce ss specifications of the object using the mCRL2 language. The concrete spec ification is built according to the implementation of the concurrent data s tructure\, while the abstract specification is linearizable by construction . Then\, linearizability can be tested by checking that their respective la beled transition systems\, generated from the mCRL2 tool\, are equivalent. This approach was applied on a number of concurrent data structures\, and i t detected both correct and faulty implementations.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 14 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/roxana-paval-modeling-and -verifying-concurrent-data-structures/ END:VEVENT BEGIN:VEVENT UID:20180123T0426Z-1516681567.5727-EO-224-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180122T154018Z LAST-MODIFIED:20180708T174625Z DTSTART;TZID=Europe/Amsterdam:20180125T124500 DTEND;TZID=Europe/Amsterdam:20180125T133000 SUMMARY: Mahmoud Talebi: Dynamic Performance Analysis of IEEE 802.15.4 Netw orks under Intermittent Wi-Fi Transmission DESCRIPTION: The coexistence of ZigBee and WiFi networks is one of the topi cs which has been investigated both empirically and mathematically. The emp irical studies often lack scalability since the networks considered are ver y small. The mathematical models on the other hand often focus on a stable network\, ignoring he dynamic interactions of unstable systems. We provide ...continue reading X-ALT-DESC;FMTTYPE=text/html:The coexistence of ZigBee and WiFi network s is one of the topics which has been investigated both empirically and mat hematically. The empirical studies often lack scalability since the network s considered are very small. The mathematical models on the other hand ofte n focus on a stable network\, ignoring he dynamic interactions of unstable systems. We provide a real-time analysis method for studying the effect of intermittent WiFi on communication in large ZigBee networks. Using our mode l we explain some of the common patterns seen in the analysis results. More over we establish that in addition to the ZigBee configuration and the clea r channel rate of WiFi\, the WiFi transmission pattern also has a high impa ct on the performance of the ZigBee network.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/mahmoud-talebi-dynamic-pe rformance-analysis-of-ieee-802-15-4-networks-under-intermittent-wi-fi-trans mission/ END:VEVENT BEGIN:VEVENT UID:20180117T0148Z-1516153713.4223-EO-213-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180116T185701Z LAST-MODIFIED:20180708T174625Z DTSTART;TZID=Europe/Amsterdam:20180118T124500 DTEND;TZID=Europe/Amsterdam:20180118T133000 SUMMARY: Wan Fokkink: Precongruence Formats with Lookahead through Modal De composition DESCRIPTION: Bloom\, Fokkink & van Glabbeek (2004) presented a method to de compose formulas from Hennessy-Milner logic with regard to a structural ope rational semantics specification. A term in the corresponding process algeb ra satisfies a Hennessy-Milner formula if and only if its subterms satisfy certain formulas\, obtained by decomposing the original formula. They used this decomposition method to ...continue reading X-ALT-DESC;FMTTYPE=text/html:Bloom\, Fokkink & van Glabbeek (2004) pres ented a method to decompose formulas from Hennessy-Milner logic with regard to a structural operational semantics specification. A term in the corresp onding process algebra satisfies a Hennessy-Milner formula if and only if i ts subterms satisfy certain formulas\, obtained by decomposing the original formula. They used this decomposition method to derive congruence formats in the realm of structural operational semantics. In this talk it is shown how this framework can be extended to specifications that include bounded l ookahead in their premises. This extension is used in the derivation of a c ongruence format for the partial trace preorder.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/wan-fokkink-precongruence -formats-with-lookahead-through-modal-decomposition/ END:VEVENT BEGIN:VEVENT UID:20180110T1111Z-1515582664.262-EO-208-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20180109T121457Z LAST-MODIFIED:20180708T174625Z DTSTART;TZID=Europe/Amsterdam:20180111T124500 DTEND;TZID=Europe/Amsterdam:20180111T133000 SUMMARY: Allan van Hulst: Completeness of Axioms for the Kleene Star under Bisimulation Equivalence: Some Progress DESCRIPTION: Robin Milner proposed a set of axioms for the Kleene star to c apture bisimulation equivalence. It is an open question whether this (or an y other) set of axioms is complete. I would like to take this opportunity t o report on some progress which has been made regarding this problem: 1) It is sufficient to rewrite ...continue reading X-ALT-DESC;FMTTYPE=text/html:Robin Milner proposed a set of axioms for the Kleene star to capture bisimulation equivalence. It is an open question whether this (or any other) set of axioms is complete. I would like to tak e this opportunity to report on some progress which has been made regarding this problem: 1) It is sufficient to rewrite terms under bisimulation to s olve this problem\, 2) I have proved completeness for all cases where the s tar nesting depth is limited to two. In particular\, the first result shows that the difficulty of this problem does not lie in deriving axiomatic equ ality\, but in finding the correct normal form under bisimilarity.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/allan-van-hulst-completen ess-of-axioms-for-the-kleene-star-under-bisimulation-equivalence-some-progr ess/ END:VEVENT BEGIN:VEVENT UID:20180105T1001Z-1515146518.0134-EO-204-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20171219T130709Z LAST-MODIFIED:20180708T174625Z DTSTART;TZID=Europe/Amsterdam:20171221T124500 DTEND;TZID=Europe/Amsterdam:20171221T133000 SUMMARY: Eelco Visser: Specification of Type Systems in Spoofax DESCRIPTION: The Spoofax language workbench supports the creation of progra mming environments for software languages using high-level declarative meta -languages for the various aspects of language definition. In this talk\, I will give a brief overview of the capabilities of Spoofax and then zoom in on its meta-language for the specification of type systems based on scope graphs ...continue reading X-ALT-DESC;FMTTYPE=text/html:The Spoofax language workbench supports th e creation of programming environments for software languages using high-le vel declarative meta-languages for the various aspects of language definiti on. In this talk\, I will give a brief overview of the capabilities of Spoo fax and then zoom in on its meta-language for the specification of type sys tems based on scope graphs for name binding and type constraints.
Lin
ks
-
http://www.metaborg.org
- A Theory of Name Resolution. ESOP 2016.
http://dx.doi.org/10.1007/978-3-662-46669-8_9
- A cons
traint language for static semantic analysis based on scope graphs. PEPM 20
16. http://doi.acm.org/10.1145/2847538.2847543
In approximately 1993 Paul Klint and cowor kers developed the ATerm library as a basic library to perform language tra nsformations. The intention from the outside was that this library would be suitable for parallel processing\, but this never materialised. Throughout the years several experiment have been done and a number of quite advanced lock and wait free algorithms have been developed that are suitable to mak e the ATerm library parallel. We review some of these algorithms and the ex cessive time to develop them. With new standard language constructs in C++1 1 it is possible to make the ATerm library parallel and we report on the cu rrent state of affairs which as it stands is not yet convincing.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/jan-friso-groote-experime nts-with-a-parallel-aterm-library-through-the-years/ END:VEVENT BEGIN:VEVENT UID:20180105T1001Z-1515146517.9998-EO-112-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20171207T090054Z LAST-MODIFIED:20180122T154534Z DTSTART;TZID=Europe/Amsterdam:20171207T124500 DTEND;TZID=Europe/Amsterdam:20171207T133000 SUMMARY: Bas Luttik: A Kleene theorem for pomset languages DESCRIPTION: We take as starting point Kleene’s well-known theorem establis hing a correspondence between regular expressions and finite automata: ever y language denoted by a regular expression is accepted by a finite automato n\, and every language accepted by a finite automaton is denoted by a regul ar expression. We extend regular expressions with a parallel construct and finite automata ...continue reading X-ALT-DESC;FMTTYPE=text/html:We take as starting point Kleene’s well-kn own theorem establishing a correspondence between regular expressions and f inite automata: every language denoted by a regular expression is accepted by a finite automaton\, and every language accepted by a finite automaton i s denoted by a regular expression. We extend regular expressions with a par allel construct and finite automata with a fork construction and establish a Kleene correspondence between the two in pomset language semantics.
This may be considered as a first step towards a Kleene theorem for Concur rent Kleene Algebra (CKA)\, an extension of Kleene Algebra with a parallel construct\, that was proposed by Hoare\, Möller\, Struth and Wehrman as a s uitable formal framework for the study of concurrent programs. CKA does req uire a different semantics satisfying the so-called exchange law between se quential and parallel composition.
The talk is based on joint work wi th Tobias Kappé\, Paul Brunet\, Alexandra Silva and Fabio Zanasi\, all from UCL.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/bas-luttik-a-kleene-theor em-for-pomset-languages/ END:VEVENT BEGIN:VEVENT UID:20180105T1001Z-1515146517.9934-EO-138-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20171130T090034Z LAST-MODIFIED:20180708T174625Z DTSTART;TZID=Europe/Amsterdam:20171130T124500 DTEND;TZID=Europe/Amsterdam:20171130T133000 SUMMARY: Sander de Putter: Evaluation of compositional model checking DESCRIPTION: Although model checking is one of the most successful approach es for the analysis and verification of the behaviour of concurrent systems \, it is plagued with the so-called state space explosion problem\; the sta te space of a concurrent system tends to increase exponentially as the numb er of parallel processes increases linearly. To combat state space explosio n ...continue reading X-ALT-DESC;FMTTYPE=text/html:Although model checking is one of the most successful approaches for the analysis and verification of the behaviour o f concurrent systems\, it is plagued with the so-called state space explosi on problem\; the state space of a concurrent system tends to increase expon entially as the number of parallel processes increases linearly. To combat state space explosion several compositional verification approaches have be en proposed such as compositional aggregation and assume-guarantee reasonin g. In their evaluation of assume guarantee reasoning Cobleigh\, Avrunin\, a nd Clarke[1] rais doubts about the effectiveness of assume-guarantee reason ing. Inspired by this work\, my current work investigates the effectiveness of compositional aggregation. In contrast to [1]\, we also aim to characte rize situations when compositional aggregation is or is not effective. In t his talk I will briefly discuss [1] and present our methodology and prelimi nary results of our evaluation of compositional aggregation.
[1] Cobl eigh\, J.M.\, Avrunin\, G.S.\, Clarke\, L.A.: Breaking up is hard to do: An evaluation of automated assume-guarantee reasoning. ACM Trans. Softw. Eng. Methodol. 17(2)\, 7:1-7:52(May 2008)
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/sander-de-putter-evaluati on-of-compositional-model-checking/ END:VEVENT BEGIN:VEVENT UID:20180105T1001Z-1515146517.9871-EO-146-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20171120T090051Z LAST-MODIFIED:20181031T192635Z DTSTART;TZID=Europe/Amsterdam:20171123T143000 DTEND;TZID=Europe/Amsterdam:20171123T153000 SUMMARY: Perry van Wesel: Formal analysis of ring networks DESCRIPTION: Systems-on-Chips rely on the correct functioning of the commun ication between their components. As these systems grow more complex\, so d o the underlying communication networks. Simulating a network does not guar antee the entire state-space is explored\, therefore formal verification me thods should be used to ensure correctness of these Networks-on-Chips. This thesis uses the MaDL modelling language ...continue reading X-ALT-DESC;FMTTYPE=text/html:Systems-on-Chips rely on the correct funct ioning of the communication between their components. As these systems grow more complex\, so do the underlying communication networks. Simulating a n etwork does not guarantee the entire state-space is explored\, therefore fo rmal verification methods should be used to ensure correctness of these Net works-on-Chips. This thesis uses the MaDL modelling language to model and v erify liveness of networks. Although MaDL is a useful tool for the verifica tion of networks\, it lacks the capability to completely model and analyse ring networks. Therefore\, MaDL is extended with a new primitive and a ring detection algorithm that cooperate to generate additional network invarian ts specifically for ring networks. These extensions are used in two case st udies to model and verify network architecture proposals from literature. T he analysed networks are TornadoNoC [10] and LIGERO [1]. In both case studi es\, the process of formally modelling these architectures reveals ambiguit y in the original proposal papers. Both architectures turn out to contain d eadlocks when modelling as close to their original proposal as possible. Th e results show formal modelling is a useful tool to eliminate ambiguity\, b ut still requires improvement to allow it to scale to larger networks. The extensions to MaDL prove to be useful when modelling certain networks\, but can still be improved to allow verification of a wider range of different networks.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 14 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/perry-van-wesel-formal-an alysis-of-ring-networks/ END:VEVENT BEGIN:VEVENT UID:20180105T1001Z-1515146517.9798-EO-142-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20171123T090033Z LAST-MODIFIED:20180708T174626Z DTSTART;TZID=Europe/Amsterdam:20171123T124500 DTEND;TZID=Europe/Amsterdam:20171123T133000 SUMMARY: Rodin Aarssen: Towards a generic framework for analyzing C++ in Ra scal DESCRIPTION: Maintenance of legacy software is often a costly task. Softwar e analysis tools can help a lot by providing insight in existing code. Howe ver\, for C++\, this tooling often doesn’t give satisfactory answers. In th is talk\, I will introduce ClaiR\, a generic C++ analysis framework I built on top of the meta-programming language Rascal. Also\, I ...continue readi ng X-ALT-DESC;FMTTYPE=text/html:Maintenance of legacy software is often a costly task. Software analysis tools can help a lot by providing insight in existing code. However\, for C++\, this tooling often doesn’t give satisfa ctory answers. In this talk\, I will introduce ClaiR\, a generic C++ analys is framework I built on top of the meta-programming language Rascal. Also\, I will discuss its future directions and show a modeling result.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/rodin-aarssen-towards-a-g eneric-framework-for-analyzing-c-in-rascal/ END:VEVENT BEGIN:VEVENT UID:20180105T1001Z-1515146517.9739-EO-145-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20171119T090005Z LAST-MODIFIED:20181031T192504Z DTSTART;TZID=Europe/Amsterdam:20171121T100000 DTEND;TZID=Europe/Amsterdam:20171121T110000 SUMMARY: Olav Bunte: Quantitative model checking on probabilistic systems u sing plmu*+ DESCRIPTION: Although most model checking logics focus on proving qualitati ve properties\, with the logic plmu*+ introduced in Mio’s PhD thesis one ca n check the probability that some behaviour happens. In this work we try to bring this logic to practice. We attempt to find intuitive meaning for plm u*+-formulas with the aim to create guidelines to create ...continue readin g X-ALT-DESC;FMTTYPE=text/html:Although most model checking logics focus on proving qualitative properties\, with the logic plmu*+ introduced in Mio 's PhD thesis one can check the probability that some behaviour happens. In this work we try to bring this logic to practice. We attempt to find intui tive meaning for plmu*+-formulas with the aim to create guidelines to creat e meaningful formulas. Also\, we give an alternative representation of a pl mu*+ model checking problem in form of an equations system and we give an a lgorithm to extract the solution from this representation\, based on the wo rk of Mader. This algorithm will be compared to an approximation algorithm by applying both on a number of use cases.
CATEGORIES:MSc Defence LOCATION:MetaForum MF 3 GEO:51.447567;5.487443 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/olav-bunte-quantitative-m odel-checking-on-probabilistic-systems-using-plmu/ END:VEVENT BEGIN:VEVENT UID:20180105T1001Z-1515146517.9669-EO-143-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20171116T090015Z LAST-MODIFIED:20180708T174626Z DTSTART;TZID=Europe/Amsterdam:20171116T124500 DTEND;TZID=Europe/Amsterdam:20171116T133000 SUMMARY: Tim Willemse: Explorations of Attribute-Based Access Control DESCRIPTION: Attribute-Based Access Control (ABAC) is emerging as the de fa cto paradigm for the specification and enforcement of access control polici es. Nonetheless\, ABAC is vulnerable to attribute hiding attacks where user s can obtain a more favourable decision by hiding some of their attributes. The extended evaluation of an ABAC policy takes such attribute hiding into account ...continue reading X-ALT-DESC;FMTTYPE=text/html:Attribute-Based Access Control (ABAC) is e merging as the de facto paradigm for the specification and enforcement of a ccess control policies. Nonetheless\, ABAC is vulnerable to attribute hidin g attacks where users can obtain a more favourable decision by hiding some of their attributes. The extended evaluation of an ABAC policy takes such a ttribute hiding into account and arguably allows to come to more precise de cisions. An extended evaluation of a given query is calculated using the ev aluation of all (sensible) queries that can be constructed from that query. Evidently\, this approach may require exploring the state space for all po ssible queries\; as such\, evaluating a query may not be particularly effic ient. In this talk we explore various techniques for computing the extended evaluation.
CATEGORIES:Colloquium LOCATION:MetaForum MF 6.131 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/tim-willemse-explorations -of-attribute-based-access-control/ END:VEVENT BEGIN:VEVENT UID:20180105T1001Z-1515146517.9567-EO-144-1@37.128.148.44 STATUS:CONFIRMED DTSTAMP:20240319T075845Z CREATED:20171102T090034Z LAST-MODIFIED:20180708T174626Z DTSTART;TZID=Europe/Amsterdam:20171102T124500 DTEND;TZID=Europe/Amsterdam:20171102T133000 SUMMARY: Omar Alzuhaibi: Automata Learning in an Engineering Context DESCRIPTION: Implementing the theory of automata learning on practical syst ems comes with many challenges\, more so when learning legacy systems. In t his talk\, I will present some of the challenges I faced while learning a l egacy component in an industrial setting at Philips Healthcare\, how I mana ged to overcome some of those challenges\, and how I ...continue reading X-ALT-DESC;FMTTYPE=text/html:Implementing the theory of automata learni ng on practical systems comes with many challenges\, more so when learning legacy systems. In this talk\, I will present some of the challenges I face d while learning a legacy component in an industrial setting at Philips Hea lthcare\, how I managed to overcome some of those challenges\, and how I pl an to solve the rest. Such challenges range from initially setting up the l earning environment to dealing with asynchronous calls and parallelism to s calability and efficiency.
CATEGORIES:Colloquium LOCATION:MetaForum MF 7.084 GEO:51.447532;5.487437 ORGANIZER;CN="Tim Willemse":MAILTO:t.a.c.willemse@gmail.com URL;VALUE=URI:https://fsa.win.tue.nl/events/event/omar-alzuhaibi-automata-l earning-in-an-engineering-context/ END:VEVENT BEGIN:VTIMEZONE TZID:Europe/Amsterdam BEGIN:STANDARD TZOFFSETFROM:+0200 TZOFFSETTO:+0100 DTSTART:20171029T010000 TZNAME:CET END:STANDARD BEGIN:DAYLIGHT TZOFFSETFROM:+0100 TZOFFSETTO:+0200 DTSTART:20180325T010000 TZNAME:CEST END:DAYLIGHT BEGIN:STANDARD TZOFFSETFROM:+0200 TZOFFSETTO:+0100 DTSTART:20181028T010000 TZNAME:CET END:STANDARD BEGIN:DAYLIGHT TZOFFSETFROM:+0100 TZOFFSETTO:+0200 DTSTART:20190331T010000 TZNAME:CEST END:DAYLIGHT BEGIN:STANDARD TZOFFSETFROM:+0200 TZOFFSETTO:+0100 DTSTART:20191027T010000 TZNAME:CET END:STANDARD BEGIN:DAYLIGHT TZOFFSETFROM:+0100 TZOFFSETTO:+0200 DTSTART:20200329T010000 TZNAME:CEST END:DAYLIGHT BEGIN:STANDARD TZOFFSETFROM:+0200 TZOFFSETTO:+0100 DTSTART:20201025T010000 TZNAME:CET END:STANDARD BEGIN:DAYLIGHT TZOFFSETFROM:+0100 TZOFFSETTO:+0200 DTSTART:20210328T010000 TZNAME:CEST END:DAYLIGHT BEGIN:STANDARD TZOFFSETFROM:+0200 TZOFFSETTO:+0100 DTSTART:20211031T010000 TZNAME:CET END:STANDARD BEGIN:DAYLIGHT TZOFFSETFROM:+0100 TZOFFSETTO:+0200 DTSTART:20220327T010000 TZNAME:CEST END:DAYLIGHT BEGIN:STANDARD TZOFFSETFROM:+0200 TZOFFSETTO:+0100 DTSTART:20221030T010000 TZNAME:CET END:STANDARD BEGIN:DAYLIGHT TZOFFSETFROM:+0100 TZOFFSETTO:+0200 DTSTART:20230326T010000 TZNAME:CEST END:DAYLIGHT BEGIN:STANDARD TZOFFSETFROM:+0200 TZOFFSETTO:+0100 DTSTART:20231029T010000 TZNAME:CET END:STANDARD END:VTIMEZONE END:VCALENDAR