Test suite minimization using particle swarm optimization

Abstract

Software engineering is a discipline which promises to produce quality software that exceeds customer expectation. In order to make these pledge realities, software testing is indispensable. More efficient and effective testing is conducted through automated testing which involves the use of automatically generated test cases. In Regression testing when the size of test suite increases running all the test cases in a test suite requires a large amount of effort and time so becomes infeasible to run all test cases. Various methods have been proposed to address these test suite minimization problem but because of its NP-completeness there no single method which produces optimum size set of test suite. In this regard we proposed a novel techniques for test suite minimization using nature inspired metaheuristic particle swarm optimization algorithm for removing the redundant test cases from the suite. We compared our technique with four benchmark reduction techniques G WSC, G, HGS and GRE based on the size of reduced set and execution cost. On the same input dataset the experimental result shows that the reduction percentage of the test suite by PSO is 55.55%, by G WSC ism22.22% and by G, HGS& GRE is 44.45%. The execution cost of the G WSC is 63, G& GRE is 72, HGS 67 and PSO is 43. Therefore, as compared with other techniques our approach showed a promising and better result.