An automated framework for verifying or refuting trace properties of extended finite state machines

Abstract

AbstractModel checkers and interactive proof assistants are both used in the assurance of critical systems. Where theorem proving involves the use of axioms and inference rules to mathematically prove defined properties, model checkers can be used to provide concrete counterexamples to refute them. Thus, the two techniques can be thought of as complementary, and it is helpful to use both in tandem to take advantage of their respective strengths. However, this requires us to translate our system model and our desired properties between the two tools which is a time-consuming and error prone process if done manually. The key contribution of this work is a set of automated tools to translate between the Isabelle/HOL proof assistant and the Symbolic Analysis Laboratory (SAL) model checker. We focus on systems specified as extended finite state machines (EFSMs) and on properties specified in linear temporal logic (LTL). We present our representations in the two tools and demonstrate the applicability of our system with respect to an academic example and two realistic case studies. This is a significant contribution to broadening the applicability of these formal approaches, since it allows two powerful verification tools to be easily used in tandem without the risk of human error.