A dash of fairness for compositional reasoning

Abstract

Proofs of progress properties often require fairness assumptions. Directly incorporating global fairness assumptions in a compositional method is difficult, given the local flavor of such reasoning. We present a fully automated local reasoning algorithm which handles fairness assumptions through a process of iterative refinement. Refinement strengthens local proofs by the addition of auxiliary shared variables which expose internal process state; it is needed as local reasoning is inherently incomplete. Experiments demonstrate that the new algorithm shows significant improvement over standard model checking. © 2010 Springer-Verlag.