Abstract:In software testing process, efficient test case generation can dramatically simplify testing, improve test efficiency and save software development costs. As an effective search algorithm, genetic algorithm has been widely applied to the study on automatic generation of test cases, and has good global search capability. However, some inherent limits of this algorithm exist, such as low optimization efficiency, premature convergence, etc. This paper proposes a modified genetic algorithm combined with tabu search algorithm, improves the select and crossover operator of genetic algorithm against the shortcomings of premature convergence, and adopt the optimal preservation strategy for improving search capabilities in the local space and the overall operating efficiency. Experiments result shows that the new algorithm has obvious advantages in efficiency and effectiveness compared with traditional genetic algorithm for test case generation.