Modular state spaces for prioritised petri nets

Abstract

Verification of complex systems specification often encounters the so-called state space explosion problem, which prevents exhaustive model-checking in many practical cases. Many techniques have been developed to counter this problem by reducing the state space, either by retaining a smaller number of relevant states, or by using a smart representation. Among the latter, modular state spaces [CP00, LP04] have turned out to be an efficient analysis technique in many cases [Pet05]. When the system uses a priority mechanism (e.g. timed systems [LP07]), there is increased coupling between the modules — preemption between modules can occur, thus disabling local events. This paper shows that the approach is still applicable even when considering dynamic priorities, i.e. priorities depending both on the transition and the current marking.