Research Specifications

Safety is one of the most important features of modern software systems, especially safety-critical systems such as nuclear power plants, which can be checked exactly by model checking. Model checking is a formal verification technique that analyzes system properties through exploring all reachable states (state space) of a model of a system. The problem of the technique is that it confronts the state space explosion in large and complex systems due to exponential memory usage. Recent researches show that a partial and intelligent exploration of the state space can be a suitable solution to overcome this problem. In this paper, we propose a two-phase model checking for safety analysis of systems specified formally through graph transformations. In the first phase, the beam-search algorithm explores the state space to a specific number of states. In case of failure of the phase, the second phase starts: in systems specified through graph transformations, the rule applied on the previous state can determine the rule that can perform on the next state. In other words, the rule on current state depends on only the rule applied to previous state, not the one on earlier states. Hence, a Markov chain (MC) is estimated to capture dependencies between the sequence of applied rules in the state space explored by the beam-search algorithm. The MC is then employed to explore the remainder of the state space intelligently. To evaluate the effectiveness of the two-phase model checking, we implement it in GROOVE, an open source toolset for designing and model checking graph transformation systems. Experimental results show that the two-phase model checking has the high speed and accuracy in comparison with the existing meta-heuristic and evolutionary techniques