SOG-Based Multi-Core LTL Model Checking

Abstract

The model checking is one of the major techniques used in the formal verification. This technique builds on an automatic procedure that takes a model M of a system and a formula φ expressing a temporal property, and decides whether the system satisfies the property (denoted by M\modelsφ). The model checking technique is based on an exhaustive exploration of the state space of the system and, thus suffers from the state space explosion problem: it can happen that the verification process stops because of lack of time or space. Among the existing solutions to tackle this problem the Symbolic Observation Graph (SOG) has been proposed as a reduced representation of the reachability graph preserving linear temporal logic properties (LTL) i.e. checking an LTL property on the SOG is equivalent to check it on the original state space. The parallel construction of the SOG could increase the speedup and scalability of model checking. In this paper, we propose a new model checking algorithm built on a parallel construction of the SOG. The SOG is adapted to allow the preservation of both state and event-based LTL formulae i.e., the atomic propositions involved in the formula to be checked could be either state-based or event-based propositions. We implemented the proposed model checking algorithm within a C++ prototype and compared our preliminary results with the state of the art model checkers.