Test Data Generation For Efsm Models Based On Tabu Search Algorithm

With the rapid development of computer software industry and the increasingly sophisticated of demand, improving the quality of software products become more and more critical, also become the focus of attention. Software testing is the most important means for the software quality assurance and become the decisive stage in the software development process. As the increasing scale of software and more complex of the structure, the testing cost is relatively increased. Automatic software testing is the major research for improving efficiency and reducing costs. However, the automatic generation of test cases is one of the most critical challenges.The software code of object-oriented software development technology has higher reuse rate, so it need strict testing to ensure software quality. However, the tester can not directly apply the process-oriented testing techniques, and followed by test models which can express a series of collection, such as UML (Unified Modeling Language) model, LTS (labeled transition system) model, EFSM (Extended Finite can State machine) model.EFSM can characterize the dynamic behavior of software system more precisely. Test case generation based on EFSM (Extended Finite State Machine Models) includes test path generation and test data generation. However, most of today’s research attention to EFSM testing focus on test path generation. In order to explore the automatic test generation, this paper presents a test data generation method with respect to the path of EFSM models. A tabu search strategy is adapted to automatically generate test data, and the key factors that affect the performance of test data generation in EFSM models are analyzed. Moreover, the test generation efficiency is compared with genetic algorithm. The experimental results show that the proposed method is promising and effective, and it is obviously superior to the genetic algorithm in the test generation for EFSM models.