基于OPTICS算法的变异体约简技术  

Mutants reduction technique research based on OPTICS algorithm

作　　者：吕学伟[1,2] 黄松[1,3] 王晔[2] 

机构地区：[1]解放军理工大学指挥信息系统学院,江苏南京210007 [2]淮阴师范学院,江苏淮安223300 [3]全军军事训练软件测评中心,江苏南京210007

出　　处：《解放军理工大学学报（自然科学版）》2016年第2期101-104,共4页Journal of PLA University of Science and Technology(Natural Science Edition)

基　　金：江苏省自然科学基金资助项目(BK20141072)

摘　　要：为了降低变异测试的计算代价,提出了一种基于OPTICS算法的变异体约简方法。首先利用OPTICS算法对变异体进行聚类,让相似的变异体在同一类簇中,然后从每个类簇中随机选取一个变异体作为代表,与其余离散的变异体组成新的变异体集合,从而达到约简变异体数量的目的。为了验证算法的效果,选择一组常用的被测程序进行实验,并和K-means算法进行比较。实验结果表明,该方法在不降低变异充分度的情况下取得了良好的约简效果,有效地减少了变异体的数量,从而降低了变异测试所产生的计算代价。To cut down high computational cost of mutant testing,a mutants reduction method was presented based on ordering points to identify the clustering structure（OPTICS）algorithm.The method classifies the similar mutants into one cluster by clustering mutants,randomly selects a mutant as representative,and then combines it with other discrete mutants to generate new mutant set.To verify the algorithm a set of common programs were tested in experiment,and the results were compared with K-means algorithm.The experimental results show that the proposed method has achieved good effects of mutants reduction,and effectively reduced the number of mutants without cutting down the mutant score of the constituted test suite,thereby reducing computational cost.

关 键 词：OPTICS算法 变异体约简 聚类技术 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]