On-the-Fly Reachability and Cycle Detection for Recursive State Machines

Searching the state space of a system using enumerative and on-the-fly depth-first traversal is an established technique for model checking finite-state systems. In this paper, we propose algorithms for on-the-fly exploration of recursive state machines, or equivalently pushdown systems, which are suited for modeling the behavior of procedural programs. We present algorithms for reachability (is a bad state reachable?) as well as for fair cycle detection (is there a reachable cycle with progress?). We also report on an implementation of these algorithms to check safety and liveness properties of recursive boolean programs, and its performance on existing benchmarks. Disciplines Computer Engineering | Computer Sciences Comments From the 11th International Conference, TACAS 2005, Held as Part of the Joint European Conferences on Theory and Practice of Software, ETAPS 2005, Edinburgh, UK, April 4-8, 2005. This conference paper is available at ScholarlyCommons: http://repository.upenn.edu/cis_papers/181 On-the-fly Reachability and Cycle Detection for Recursive State Machines ? Rajeev Alur, Swarat Chaudhuri, Kousha Etessami, and P. Madhusudan 1 University of Pennsylvania, USA 2 University of Edinburgh, UK 3 University of Illinois at Urbana-Champaign, USA Abstract. Searching the state space of a system using enumerative Searching the state space of a system using enumerative and on-the-fly depth-first traversal is an established technique for model checking finite-state systems. In this paper, we propose algorithms for on-the-fly exploration of recursive state machines, or equivalently pushdown systems, which are suited for modeling the behavior of procedural programs. We present algorithms for reachability (is a bad state reachable?) as well as for fair cycle detection (is there a reachable cycle with progress?). We also report on an implementation of these algorithms to check safety and liveness properties of recursive boolean programs, and its performance on existing benchmarks.