Bounded model checking of graph transformation systems via SMT solving

Abstract

Bounded model checking (BMC) complements classical model checking by an efficient technique for checking error-freedom of bounded system paths. Usually, BMC approaches reduce the verification problem to propositional satisfiability. With the recent advances in SAT solving, this has proven to be a fast analysis. In this paper we develop a bounded model checking technique for graph transformation systems. Graph transformation systems (GTSs) provide an intuitive, visual way of specifying system models and their structural changes. An analysis of such models - however - remains difficult since GTSs often give rise to infinite state spaces. In our BMC technique we use first-order instead of propositional logic for encoding complex graph structures and rules. Today's off-the-shelf SMT solvers can then readily be employed for satisfiability solving. The encoding heavily employs the concept of uninterpreted function symbols for representing edge labels. We have proven soundness of the encoding and report on experiments with different case studies. © 2013 IFIP International Federation for Information Processing.