Detecting Local Software Issues Using NSGA Multi-optimization
摘要
The primary goal of software testing is to identify and fix bugs. If a problem exists, it must be identified, tracked down, and fixed. As the adjacent tasks, fault detection and localization are difficult to merge. When it comes to fault localization, information helps to locate defects precisely, but when it comes to fault detection, test information is needed to find issues early. However, one thing that is necessary for both of these things is test information. Combining fault detection and localization becomes much easier with this prerequisite. This paper proposes a multi-objective approach that utilizes code and mutation coverage to provide a condensed test suite capable of identifying and pinpointing faults. The NSGA-II method was used for the optimization of the test cases. The findings from the analysis of the projects in the Defects4j repository demonstrate that the suggested approach is capable of producing smaller test suites that can effectively detect 94.16% of defects and accurately pinpoint the location of these discovered defects, with a fault localization score that is comparable to that of the whole test suite. They provided scores for problem identification and localization; the normal reduction percentage suite size is 77% which is an excellent reduction percentage.