Maximilian Köhl: Most-Specific Verdictors for Imperfectly Observable Systems

Event Details

The observable behavior of a system usually carries crucial information about its internal state, properties, and potential future behaviors. In this talk, I present a generic automata-theoretic synthesis approach to obtain most-specific verdicts from imperfect observations of an ongoing run of a system. Verdicts can be elements of any join-semilattice ordered by specificity. I show that our approach covers existing work on runtime monitoring and fault diagnosis, where verdicts indicate the satisfaction or violation of properties, or the occurrence of faults, respectively. As an entirely novel application, I present configuration monitors, yielding verdicts about possible configurations of a system. Furthermore, I present empirical results obtained by configuration monitoring on well-established configurable systems community benchmarks. If there is enough time, I might also give a sneak peek on how the approach generalizes to the real-time setting, where the timing of events is often crucial for monitoring and diagnosis, but its precise assessment is hindered by timing imprecisions.