Application of Improved Harmony Search Algorithm in Test Case Selection

Abstract

Regressiontest is an effective means to ensure the quality of software. But the testefficiency will become lower and lower as the test case library becomesincreasingly large in the test. Therefore, in order to improve the efficiencyand quality of regression test, it is necessary to implement the test casesafety reduction. The optimization selection of the test case is an effectivemeans of the test case safety reduction. In order to solve the problem of testcases optimization selection, this paper proposes a test case selection methodbased on the improved harmony search algorithm. Specifically, the researcheradds an excellent harmony element library based on the classical harmony searchalgorithm for storing the searched excellent test case; meanwhile, he alsoimproves the way of generating the new harmony toallow these excellent test cases to be retained with a certain probability. Inthe search process, the researcher makes the adaptive adjustment of thealgorithm parameter BW in a non-linear incremental manner, improving the globalsearch capability of the algorithm. In addition, the excellence protectionstrategies are also increased to prevent against the degradation in theoptimization search process. In order to verify the feasibility of the improvedalgorithm, the researcher writes four programs: the genetic algorithm,differential evolution algorithm, classical harmony algorithm and improvedalgorithm, and conducted the comparative experiment. The experimental resultshows that the improved algorithm significantly improves the optimizationsearch performance compared with the classical harmony search algorithm and hasthe better performance compared with the genetic algorithm and differentialevolution algorithm.