| Abstrakt: |
Testing is essential for the software’s success, but despite this, it is a time and resource-consuming activity. Therefore, researchers and practitioners continuously try to improve software testing automation to maximize test coverage and make it fast and reliable. Meta-heuristics are high-level frameworks that many researchers have used to generate test data for software testing. Maximizing test coverage would also indirectly help to find vulnerabilities in the code. In this paper, we have implemented an improved hybrid metaheuristic algorithm to generate test cases, utilizing particle swarm optimization (PSO) and genetic algorithm (GA) for path coverage testing criterion. The used fitness function is the combination of branch distance, approximation level and path distance. The proposed approach is a hybrid Particle Swarm Optimization and Genetic Algorithm (PSO-GA). We compared the meta-heuristics GA, PSO and Hybrid PSO-GA algorithm with different fitness functions. Moreover, the experimental result shows that the hybrid algorithm improves outcomes compared to GA and PSO for the combined fitness functions. This approach demonstrates noteworthy efficacy in addressing security vulnerabilities during testing, particularly due to its emphasis on comprehensive path testing. This methodology has yielded significant outcomes in the realm of security testing, highlighting its potential for practical application and exploration in research. Furthermore, more meta-heuristics can be incorporated into the hybrid approach. |
Full text from SpringerLink