Optimum Selection of PSO Parameters for Automated Test Data Generation of Software Programs

The proposed research work implemented and fine-tuned meta-heuristic based search algorithms namely PSO algorithm for automatic test case generation using path testing criterion. The three parameters namely size of population (NIND), inertia weight (W), social learning rate (C1), and cognitive learning rate (C2) have been chosen for setting PSO algorithm for test data generation. For input generation, symbolic execution method has been used in which first, target path is selected from Control Flow Graph (CFG) of Software under Test (SUT) and then inputs are generated using search algorithms which can evaluate composite predicate corresponding to the target path true. We have experimented on two real world programs showing the applicability of these techniques in genuine testing environment. The algorithm is implemented using MATLAB programming environment. The performance of the algorithms is measured using average test cases generated per path (ATCPP) and average percentage coverage (APC) metrics.