Partial Order Reduction in Directed Model Checking

Abstract

Partial order reduction is a very succesful technique for avoiding the state explosion problem that is inherent to explicit state model checking of asynchronous concurrent systems. It exploits the commutativity of concurrently executed transitions in interleaved system runs in order to reduce the size of the explored state space. Directed model checking on the other hand addresses the state explosion problem by using guided search techniques during state space exploration. As a consequence, shorter errors trails are found and less search effort is required than when using standard depth-first or breadth-first search. We analyze how to combine directed model checking with partial order reduction methods and give experimental results on how the combination of both techniques performs.