Efficient Safety Verification of Discrete Event Systems

Paper i proceeding, 2023

Formal verification of safety properties is critical in many application areas. In this paper a survey of the most common and efficient methods is given. The different methods are compared for some typical scalable examples, both modeled by transition systems and Petri nets. The methods are based on symbolic computations and logic representations of both sets and transition relations. Reachability analysis is performed not only from the initial states to any forbidden states, but also in the opposite backward direction. The reason is that a system is also safe when no initial state is backward reachable from the set of forbidden states. For safe systems the number of backward reachable states from forbidden states is often much less than the forward reachable states from the initial states. The backward strategy is shown to be very efficient and often even better than the most efficient methods that are specialized on safety analysis.