基于MC/DC的回归测试数据进化生成  

Evolutionary Generation of the Regression Test Tata Based on MC/DC

作　　者：韩莹[1] 罗扬[1] 吴取劲[1] 余童兰[1] 

机构地区：[1]南华大学计算机科学与技术学院,湖南衡阳421001

出　　处：《南华大学学报（自然科学版）》2016年第1期55-60,共6页Journal of University of South China：Science and Technology

基　　金：湖南省科技厅自然科学省市联合基金资助项目(13JJ9008);湖南省教育厅科研基金资助项目(10C1144)

摘　　要：在合理利用已有测试数据形成优势初始种群的前提下采用遗传算法自动生成回归测试数据是软件测试研究的一个热点.本文通过在已有测试数据的基础上依据MC/DC准则演进增补部分用例提升MC/DC覆盖率.首先,通过记录每个已有测试数据覆盖的条件组合确定要增补用例的目标条件组合,其次,根据适应度函数从已有测试数据中筛选出部分数据作为初始种群,再次,根据已筛选的部分初始种群所覆盖的条件组合与目标条件组合确定遗传操作分量,最后,演进并判定提取目标数据.理论与实验表明,该方法可以提高回归测试数据生成效率及代码覆盖率.Automatically generating regression test data using genetic algorithms is a hotspot in the field of software testing, the premise of which is to appropriately utilize the existing test data to form the initial population.In this paper,we added some test cases to the existing test data based on the MC/DC coverage criteria to improve the MC/DC coverage.By record- ing the condition combination coverage of each existing test data, we determined the target condition combinations for supplementation.And according to the fitness function we estab- lished the initial population.After that,based on the condition combinations by which the se- lected initial population covered and the target condition combinations we determined the ge-netic components.Finally,through evolution we judged and extracted the target theoretical and experimental results show the benefits of this method to efficiency of both the regression test data generation and code coverage. data.Both the improve

关 键 词：回归测试 遗传算法 适应度函数 初始种群 软件测试 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]