Pdmc ’ 05

Enumerative distributed reachability analysis based on a simple partitioning of the state space has become a well known technique implemented in numerous model checking tools. The timed automata model checker UPPAAL is no exception to this rule, and for over 5 years a distributed version of UPPAAL has been used to analyse large timed automata models. In this talk we address two fundamental issues with the current approach. Memory usage and CPU usage is tightly coupled. In particular, this limits the freedom to utilize non-homogeneous setups such as mixed clusters or clusters of clusters (e.g. connected by GRID infrastructure). Our proposed solution also solves the load instability we first observed in early versions of distributed UPPAAL and that has since been observed in other distributed model checkers. Distributed reachability analysis is tightly coupled with breadth first search of the state space. For this reason, extending the algorithm from pure reachability analysis to LTL or TCTL model checking is difficult, as current on-the-fly approaches for timed automata are based on a depth-first search. We present an idea that allows any fixed point computation, including LTL or TCTL model checking, to be easily distributed. Preprint submitted to Elsevier Preprint 13 June 2005 6 PDMC 2005 Preliminary Version How to Order Vertices for Distributed LTL Model-Checking Based on Accepting Predecessors L. Brim , I. Černá , P. Moravec, J. Šimša Faculty of Informatics, Masaryk University, Brno, Czech Republic Abstract Distributed automata-based LTL model-checking relies on algorithms for finding accepting cycles in a Büchi automaton. The approach to distributed accepting cycle detection as presented in [9] is based on maximal accepting predecessors. The ordering of accepting states (hence the maximality) is one of the main factors affecting the overall complexity of model-checking as an imperfect ordering can enforce numerous re-explorations of the automaton. This paper addresses the problem of finding an optimal ordering, proves its hardness, and gives several heuristics for finding an optimal ordering in the distributed environment. We compare the heuristics both theoretically and experimentally to find out which of these work well.Distributed automata-based LTL model-checking relies on algorithms for finding accepting cycles in a Büchi automaton. The approach to distributed accepting cycle detection as presented in [9] is based on maximal accepting predecessors. The ordering of accepting states (hence the maximality) is one of the main factors affecting the overall complexity of model-checking as an imperfect ordering can enforce numerous re-explorations of the automaton. This paper addresses the problem of finding an optimal ordering, proves its hardness, and gives several heuristics for finding an optimal ordering in the distributed environment. We compare the heuristics both theoretically and experimentally to find out which of these work well.