Seminar 28 Sept 15:15, 2012

Verification and Synthesis for Parametric Markov Chains

Abstract: In this thesis we present a new approach for model checking Parametric Markov Chains (PMCs). In PMCs certain transition probabilities are left open by introducing parameters instead of concrete values. Model checking PMCs for unbounded reachability probabilities then yields a rational function as a probability bound for reaching the set of target states. These rational functions can be instantiated for a range of parameter values or analyzed further. Our algorithm extends an existing approach for model checking DTMCs and is implemented in the tool COMICS. Further, we developed our own data structure to simplify rational functions more efficiently. We demonstrated the applicability of our methods on a number of case studies.