Effective Lightweight Software Fault Localization based on Test Suite Optimization

Abstract

Automated software fault localization techniques aid developers in program debugging by identifying the probable locations of faults in a program with minimum human intervention. As software is growing in complexity and scale today, increasing the efficiency of fault localization techniques is very essential in order to reduce the overall software development cost. The effectiveness of the test suites used in the fault localization process has a significant impact on the efficiency of the process. Previous studies, on the other hand, have placed less focus on the adequacy of test suites for the fault localization process. We apply optimized test suites in this paper to improve the performance of software fault localization in a single-fault scenario. For our experiments, we use spectrum-based fault localization (SBFL) techniques. Because of its minimal computing overhead and scalability, spectrum-based fault localization is a popular, efficient, and yet lightweight fault localization technique. To optimize the test suite, we employ a heuristic that asserts that if a faulty statement is executed by a passing test case, that test case will have a negative impact on fault localization performance. In contrast, if a passing test case does not execute the faulty statement, the faulty statement's suspiciousness increases, which has a positive impact on fault localization performance. The test suite optimization approach used in this paper significantly improves fault localization performance, as demonstrated by our experiments. The results show that the proposed method efficiently reduces the number of statements examined by about 84.94 percent on average.