May/Must Abstraction-Based Software Model Checking for Sound Verification and Falsification

Abstract

Three-valued models, in which properties of a system are either true, false or unknown, have recently been advocated as a better representation for reactive program abstractions generated by automatic techniques such as predicate abstraction. Indeed, for the same cost, model checking three-valued abstractions, also called may/must abstractions, can be used to both prove and disprove any temporal-logic property, whereas traditional conservative abstractions can only prove universal properties. Also, verification results can be more precise with generalized model checking, which checks whether there exists a concretization of an abstraction satisfying a temporal-logic formula. Generalized model checking generalizes both model checking (when the model is complete) and satisfiability (when everything in the model is unknown), probably the two most studied problems related to temporal logic and verification. This paper presents an introduction to the main ideas behind this framework, namely models for three-valued abstractions, completeness preorders to measure the level of completeness of such models, three-valued temporal logics and generalized model checking. It also discusses algorithms and complexity bounds for three-valued model checking and generalized model-checking for various temporal logics. Finally, it discusses applications to program verification via automatic abstraction.