基于差分进化算法的组合测试用例集生成

Generate combination test cases with differential evolution algorithm

为了减小所生成的组合测试用例集规模,提出了一种基于one-test-at-a-time策略的差分进化算法来求解该问题的方案。通过实验,比较了不同变异方式的优化性能,考察了各个算法参数对优化性能的影响,并与其他常用方法进行了对比。实验结果表明,与其他常用的组合测试用例生成方法相比,基于差分进化算法的生成方法能够生成规模更小的组合测试用例集。并且组合数越多,该算法的优势就越明显。

For the same combined strength, a test suite with smaller scale is more superior. According to the characteristics of the problem, this paper proposed an approach based on one-test-at-a-time strategy differential evolution algorithm to solve the problem. By experimenting, it compared the optimum performance of different varying modes, observed and studied the effects of various parameters on the optimization algorithm performance. Comparative experiments with other commonly used methods show that the scale of combinatorial test suite generated by differential evolution algorithm is smaller.