Abstract: Applying GA in automatic test cases generation, followed is the main work: (1)Using control flow path table to analyze the relation between test cases and the executing path; (2)Implementing the GA for automatic test cases generating which is path-oriented, and proving this method can accomplish the equivalent-class subdividing in all the test cases set; (3)In the point of statistics to analyze the error-detecting capability of this method; (4)Doing experiment for validating, which shows a far better performance than a random way.

Key words: Automatic software test, Genetic algorithm, Self-adaptive, Test case

摘要： 采用遗传算法用于自动生成测试用例的设计，算法面向路径测试作了下列工作：(1)以控制路径表CFDPATH_T为基础分析测试用例与执行路径的关系；(2)设计面向路径的遗传算法实现测试用例的自动生成，证明了该方法能实现测试用例空间上对路径的等价类划分；(3)从统计角度对该方法的错误检测能力进行了量化分析；(4)在该算法的基础上设计了试验和结果分析，结果表示该算法较随机方法有更好的性能。

关键词: 自动软件测试, 遗传算法, 自适应, 测试用例