2008 | OriginalPaper | Buchkapitel
Approximation Refinement for Interpolation-Based Model Checking
verfasst von : Vijay D’Silva, Mitra Purandare, Daniel Kroening
Erschienen in: Verification, Model Checking, and Abstract Interpretation
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Model checking using Craig interpolants provides an effective method for computing an over-approximation of the set of reachable states using a SAT solver. This method requires proofs of unsatisfiability from the SAT solver to progress. If an over-approximation leads to a satisfiable formula, the computation restarts using more constraints and the previously computed approximation is not reused. Though the new formula eliminates spurious counterexamples of a certain length, there is no guarantee that the subsequent approximation is better than the one previously computed. We take an abstract, approximation-oriented view of interpolation based model checking. We study counterexample-free approximations, which are neither over- nor under-approximations of the set of reachable states but still contain enough information to conclude if counterexamples exist. Using such approximations, we devise a model checking algorithm for approximation refinement and discuss a preliminary implementation of this technique on some hardware benchmarks.