Symbolic Partial Order Reduction for Rule Based Transition Systems

Abstract

Partial order (PO) reduction methods are widely employed to combat state explosion during model-checking. We develop a partial order reduction algorithm for rule-based languages such as Murphi [4] based on the observation that for finite-state systems, independence conditions used for PO reduction can be encoded as boolean propositions and checked using SAT methods. Comparisons against static-analysis based PO reduction algorithms have yielded encouraging results.