An Improvement in Partial Order Reduction Using Behavioral Analysis
Abstract
Efficacy of partial order reduction in reducing state space relies on adequate extraction of the independence relation among possible behaviors. However, traditional approaches by statically analyzing system model structures are often not able to reveal enough independence for reduction. To address such a problem, this paper presents a behavioral analysis approach that uses a compositional reachability analysis method to generate the over-approximate local state spaces for all modules in a system where a much more precise independence relation can be extracted for partial order reduction. Compared to the static analysis approaches, significantly higher reduction on complexity can be seen in a number of non-trivial examples, and as a consequence, dramatically less time and memory are required to finish these examples.
Hao Zheng
University of South Florida, Associate Professor
