محلی سازی گسل های نرم افزاری متعدد بر اساس الگوریتم تکاملی

During software debugging, a significant amount of effort is required for programmers to identify the root cause of manifested failures. Various spectrum-based fault localization techniques have been proposed to automate the procedure. However, most of the existing fault localization approaches do not consider the fact that programs tend to have multiple faults. Considering faults in isolation results in less accurate analysis. In this paper, we propose a flexible framework called FSMFL for localizing multiple faults simultaneously based on genetic algorithms with simulated annealing. FSMFL can be easily extended by different fitness functions for the purpose of localizing multiple faults simultaneously. We have implemented a prototype and conducted extensive experiments to compare FSMFL against existing spectrum based fault localization approaches. The experimental results show that FSMFL is competitive in single-fault localization and superior in multi-fault localization.