Automatic Symmetry Detection for Model Checking Using Computational Group Theory

We present an automatic technique for the detection of structural symmetry in a model directly from its Promela specification. Our approach involves finding the static channel diagram of the model, a graphical representation of channel-based system communication; computing the group of symmetries of this diagram; and computing the largest possible subgroup of these symmetries which induce automorphisms of the underlying model. We describe a tool, SymmExtractor, which, for a given model and LTL property, uses our approach to find a group of symmetries of the model which preserve the property. This group can then be used for symmetry reduction during model checking using existing quotient-based methods. Unlike previous approaches, our method can detect arbitrary structural symmetries arising from the communication structure of the model.