Compositional SCC Analysis for Language Emptiness

We propose a refinement approach to language emptiness, which is based on the enumeration and the successive refinements of SCCs on over-approximations of the exact system. Our algorithm is compositional: It performs as putation as possible on the abstract systems, and prunes uninteresting part of the search space as early as possible. It decomposes the state space disjunctively so that each state subset can be checked in isolation to decide language emptiness for the given system. We prove that the strength of an SCC or a set of SCCs decreases monotonically with composition. This allows us to deploy the proper model checking algorithms according to the strength of the SCC at hand. We also propose to use the approximate distance of a fair cycle from the initial states to guide the search. Experimental studies on a set of LTL model checking problems prove the effectiveness of our method.

• Type: Journal
• Year: 2006
• Where: FMSD
• Authors: Chao Wang, Roderick Bloem, Gary D. Hachtel, Kavita Ravi, Fabio Somenzi
• Keywords: Abstract Systems, FMSD 2006, Refinement Approach, Successive Refinements