一种变异测试中冗余变异体的寻找方法

发布时间：2018-04-25 17:06

  本文选题：变异测试 + 变异算子　； 参考：《电子学报》2017年08期

【摘要】：变异测试是一种有效的基于故障的测试方法,但大量冗余变异体所带来的昂贵的测试成本问题,阻碍了它在实际工程开发中的应用.为解决该问题,本文针对程序中的顺序语句所产生的变异体,基于故障的可达-感染-传播模型,提出了使用区间抽象域来表示程序状态,通过区间运算判断变异体之间冗余关系的算法;针对程序中的条件语句,基于谓词故障层级,分别给出了面向简单谓词和复合谓词的冗余变异体选择算法.并对这两种算法对冗余变异体的判定效果进行了分析,最后给出了在分层抽样背景下,非冗余变异体生成的约束边界条件.对Siemens和开源项目等共8个工程进行了实验,并与随机选择法进行了对比.结果表明,本文所提方法在减少变异测试时间成本的同时,可以保持较高的变异得分.
[Abstract]:Mutation testing is an effective fault based test method, but the high cost of testing caused by a large number of redundant variants hinders its application in practical engineering development. In order to solve this problem, based on the reachability-infection-propagation model of fault, an interval abstract domain is proposed to represent the state of the program. Based on the fault level of predicates, the algorithm of selecting redundant variants for simple predicates and compound predicates is presented, which is based on the conditional statements in the program. The effect of these two algorithms on the determination of redundant variants is analyzed. Finally, the constrained boundary conditions for the generation of non-redundant variants are given under the background of stratified sampling. Eight projects, such as Siemens and open source projects, are tested and compared with the random selection method. The results show that the proposed method can reduce the cost of mutation test time and maintain a high variation score.
【作者单位】： 北京邮电大学网络与交换技术国家重点实验室;桂林电子科技大学广西云计算与大数据协同创新中心;桂林电子科技大学广西高校云计算与复杂系统重点实验室;长春汽车工业高等专科学校;
【基金】：国家自然科学基金(No.91318301,No.61202080) 广西云计算与大数据协同创新中心、广西高校云计算与复杂系统重点实验室资助(No.YD16508)
【分类号】：TP311.53
，

本文编号：1802208