基于自组织映射神经网络的变异体约简方法  

Mutant reduction method based on self-organizing map neural network

作　　者：王曙燕 高雨 WANG Shuyan;GAO Yu(School of Computer Science and Technology,Xi’an University of Posts and Telecommunications,Xi’an 710121,China;Xi’an Engineering Research Center of IOT Intelligent Information Collection and Processing,Xi’an 710121,China)

机构地区：[1]西安邮电大学计算机学院,陕西西安710121 [2]西安市物联网智能信息采集与处理工程研究中心,陕西西安710121

出　　处：《西安邮电大学学报》2023年第5期50-55,共6页Journal of Xi’an University of Posts and Telecommunications

基　　金：陕西省重点研发计划项目(2020GY-210)。

摘　　要：针对变异测试中产生大量变异体导致变异测试成本过高的问题,提出一种基于自组织映射神经网络的变异体约简方法。利用弱变异转换法获得变异体杀死矩阵,将其作为变异体的特征数据,使用自组织映射神经网络对变异体聚类,并将相似的变异体放在一类簇中,根据变异体的杀死度从每类簇中选择最难杀死的变异体组成新的变异体集合,从而约简变异体的数量。测试结果表明,所提方法在保证变异测试有效性不受影响的同时可以约简平均80%的变异体,降低了变异测试成本。In view of the problem that the mutation testing cost is too high brought by a large number of mutants,a mutant reduction method based on self-organizing map neural network is proposed.The mutant killing matrix is obtained based on the weak mutation testing transformation.The killing matrix is used as the feature data of the mutants,which are clustered by the self-organizing map neural network.Mutants with similar characteristics would be placed in the same cluster.Mutants with the lowest killing degree are selected from each cluster to form a new set of mutants,so as to reduce the number of mutants.Experimental results show that the proposed method can reduce 80%mutants on average without affecting the mutation testing effectiveness,which reduces the cost of the mutation testing.

关 键 词：软件测试 自组织映射神经网络 变异测试 变异分支 变异体约简 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]