基于混合遗传算法的回归测试用例集最小化研究  

Research on regression test suite minimization based on hybrid genetic algorithm

作　　者：邓秋辉 罗小华[1] DENG Qiu-hui;LUO Xiao-hua(Institute of Very Large Scale Integrated Circuit Design,Zhejiang University,Hangzhou 310007,China)

机构地区：[1]浙江大学超大规模集成电路设计研究所,浙江杭州310007

出　　处：《传感器与微系统》2018年第7期11-14,共4页Transducer and Microsystem Technologies

基　　金：浙江省自然科学基金资助项目(LY15F040001)

摘　　要：为缩减测试用例规模,降低回归测试成本,将遗传算法和贪心算法相结合,提出了一种混合遗传算法用于解决测试用例最小化问题。算法对标准遗传算法中的选择、交叉和变异操作进行改善,提高了算法的全局寻优能力。同时,利用贪心算法处理可行解和不可行解,提高了算法的局部寻优能力。实验结果表明:与标准遗传算法相比,在保证测试完备性的前提下,混合遗传算法能够得到更优的缩减效果和更快的收敛速度。In order to reduce scale of test cases and to reduce the cost of regression test, combine genetic algorithm with greedy algorithm,propose a hybrid genetic algorithm to solve the problem of test case minimization. Selection ,crossover and mutation operation in standard genetic algorithm are improved by this algorithm, which effectively improve global optimizing ability of the algorithm. Meanwhile, greedy algorithm is used to deal with feasible solutions and infeasible solutions, which can improve local optimizing ability of genetic algorithm. Experimental results show that compared with the standard genetic algorithm, the hybrid algorithm can obtain better reduction effect and faster convergence speed under the premise of guaranteeing the completeness of the test.

关 键 词：回归测试 测试用例集最小化 遗传算法 贪心算法 

分 类 号：TN47[电子电信—微电子学与固体电子学]