Symbolic Model Checking with Fewer Fixpoint Computations

Abstract

Symbolic model checking, smc, is a decision procedure that verifies that some finite-state structure is a model for a formula of Computation Tree Logic (CTL). smc is based on fixpoint computations. Unfortunately, as the size of a structure grows exponentially with the number of state components, smc is not always powerful enough to handle realistic problems.We first show that a subset of CTL formulas can be checked by testing simple sufficient conditions, that do not require any fixpoint computation. Based on these observations, we identify a second, larger, subset of CTL that can by verified with fewer fixpoint computations than smc. We propose a model checking algorithm for CTL that tests the identified sufficient conditions whenever possible and falls back to smc otherwise. In the best (resp. worst) case, the complexity of this algorithm is exponentially better (resp. the same) in terms of state components than that of smc.KeywordsModel CheckTemporal LogicCient ConditionKripke StructureComputation Tree LogicThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.