基于自学习控制算法优化软件测试用例的研究

Research on optimal software testing case based on self learning control algorithm

研究了软件参数变化条件下,在回归测试中以最快速度修复软件缺陷为目标的软件优化问题,将软件测试过程转化为一个时变系统控制过程,给出了软件测试状态转移矩阵模型。运用学习控制方法,通过二维变因子自学习策略获得软件测试最优测试用例,优化软件测试。仿真结果表明,给出的学习策略优于随机测试和马尔可夫控制策略,对应地检测与排除同样软件缺陷,该控制策略能显著减少回归测试次数,降低测试成本。

This paper demonstrates an approach to optimize software testing by rapid fixing software bug with given software parameter uncertainty during regressive testing process.Taking software testing process into time-varied system control problem,a state transform matrix model is presented.The two dimension variable factor self learning strategy is used to get optimized test case.Simulation results show that the learning control strategy is better than random testing and Markov testing strategy,and can significantly reduce regressive test numbers and save test cost.