Research On Test Case Generation Based On Improved PSO Algorithm

Automatic generation of test case is an important part of software testing automation. It can improve the efficiency of soft testing, reduce the costs of software development and ensure software quality. Among the methods of path test case generation, the traditional techniques (random method, static method and dynamic method, etc.) are inefficient, complex and inadequate to solve the problem. Although genetic algorithm has the superiority in test case generation, it need to encode and decode. The complex operation leads the efficiency decreased. How quickly and effectively generate the test case has a very important theoretical and practical significance.PSO (Particle Swarm Optimization, PSO) has a fast convergence speed and easy to use. It is a new method to solve the problem that automatic generation the path test case. But the local search ability of PSO is poor, and the search accuracy is not high, easy to fall into the local minimum.The paper improves the algorithm and proposes a method of test case automatic generation based on adaptive particle swarm optimization algorithm with chaos local search (APSOCLS). The major work of this paper includes:Firstly, the adaptive strategies: Inertia weight is a key to affect the convergence speed and performance of the algorithm. In the paper, the inertia weight of the particle was adjusted adaptive based the convergence degree of the swarm and the fitness of the particle. The diversity of inertia weight makes a balance between the global exploration and local improvement, so it can improve the convergence speed and accuracy of the algorithm.Secondly, adding chaotic local search (CLS): The local search ability of the PSO is poor, when to the near of the local minimum, the convergence speed is become slowly and easy to appear the vibration. So we add the chaotic local search in PSO to solve this problem. In order to improve the speed of the chaotic search, we adopt a Tent map that map distribution more uniform, and improve for its own existence to the fixed point problem.Thirdly, Design the algorithm for path test case generation, to improve the efficiency of generating test case.Finally, we verified that the model for the test case generation. Experiment from number of iterations and the running time were measured and with the existing PSO algorithms and immune genetic algorithm were compared, experimental results show that adaptive particle swarm optimization algorithm with chaos local can effectively improve the efficiency of the test case generation. Based on the research, we development a practical tool using VB. This tool can automatic generate test case for the specified path, and has a certain practicality.