Abstract: Genetic algorithm is currently an adaptive searching algorithm which gives a new solution to generate test cases. However, when it is used to generate test cases, the randomness of recombination and mutation makes the diversity of the population easily destroyed and leads to the search space getting smaller, finally results in the algorithm finding a wrong local optimum. Moreover, recombining and mutating freely makes the searching efficiency very low. This paper introduces an improved method, with a mutation control strategy and an optimal control strategy added, which effectively improves the searching capability and gets optimal solution capability of the algorithm.

Key words: genetic algorithm, test case, mutation control strategy, optimal control strategy, optimal solution