An Optimization Strategy to Feature Models’ Verification by Eliminating Verification-Irrelevant Features and Constraints

Abstract

Feature models provide an effective approach to requirements reuse. One important problem related to feature models is the verification problem, which is NP-complete in theory. The existing approaches to feature models’ verification mostly focus on how to automate the verification of feature models using third-party’s tools, while these tools are usually designed to resolve general kinds of problems. However, by simply using these third-party’s tools, large-sized feature models still can hardly be verified within acceptable time. We argue that, to improve the efficiency of verification, the problem itself should be at first optimized. In this paper, we propose an optimization strategy to feature models’ verification, in which, verification-irrelevant features and constraints are eliminated from feature models and the problem size of verification is therefore reduced. We prove the correctness of this strategy, while experiments show the effectiveness of this strategy.