基于遗传算法和程序依赖性分析的测试数据生成方法

he software testing based on branch coverage is effective to detect defects in the program. For making the branch coverage more simple and precise, a novel approach to generating test data for branch coverage is presented. This approach consists two parts, first, utilizing control dependence analysis to reduce the number of testing requirements, and the coverage of the remaining branches can guarantee the coverage of all test requirements is achieved; second, the path constraints of the target branch is obtained by using program dependence analysis on it; and then transforming the path constraints to appropriate fitness function for the target branch, and the genetic algorithm is used to generate multiple test data. By analyzing the proposed approach, it could be known that it can greatly reduce the number of testing requirements and improve the fitness function.