An Intelligent Method Based on State Space Search for Automatic Test Case Generation

摘要

Search-Based Software Testing reformulates testing as search problems so that test case generation can be automated by some chosen search algorithms. This paper reformulates path-oriented test case generation as a state space search problem and proposes an intelligent method Best-First-Search Branch & Bound to solve it, utilizing the algorithms of Branch & Bound and Backtrack to search the space of potential test cases and adopting bisection to lower the bounds of the search space. We also propose an optimization method by removing irrelevant variables. Experiments show that the proposed search method generates test cases with promising performance and outperforms some MetaHeuristic Search algorithms.