Static Partial Order Reductions for Probabilistic Systems

The present Master’s thesis seeks the development and analysis of static partial order reduction techniques for the models of probabilistic systems. The properties of those systems can be verified via model checking technique. Model checking suffers from the problem known as State Space Explosion, which can make the verification process intractable. Partial order reductions are aimed at alleviating that problem. As an outcome of the work carried out for the elaboration of current thesis, two new static partial order techniques, named Naïve SPOR and Reachability-Aware SPOR, were defined. A recommendation of the situations in which each of them should be used is provided. The latter achieves a better reduction than the former when the system to be verified is not probabilistic or when the property to be checked can be expressed in a linear temporal logic. The software tool known as LiQuor model checker was extended in order to be able to execute both reduction techniques. Those techniques were utilized for the reduction of the models of some classical concurrent systems. Several properties of the reduced and unreduced models were verified using symbolic and explicit model checking techniques. As a result of the analysis over the experiments, it is concluded that static partial order techniques should be more conveniently used in combination with symbolic model checking than with explicit model checking.