Fault Localization Based on Hybrid Genetic Simulated Annealing Algorithm

Abstract

Software testing is an important stage in the software development process, which is the key to ensure software quality and improve software reliability. Software fault localization is the most important part of software testing. In this paper, the fault localization problem is modeled as a combinatorial optimization problem, using the function call path as a starting point. A heuristic search algorithm based on hybrid genetic simulated annealing algorithm is used to locate software defects. Experimental results show that the fault localization method, which combines genetic algorithm, simulated annealing algorithm and function correlation analysis method, has a good effect on single fault localization and multi-fault localization. It greatly reduces the requirement of test case coverage and the burden of the testers, and improves the effect of fault localization.