基于控制流图支配树的测试数据灰度编码生成  

Gray encoding evolutionary with control flow dominating tree for test data generation

作　　者：单平平[1] 陈可[1] 许豪[1] 

机构地区：[1]南阳理工学院

出　　处：《计算机应用研究》2017年第3期756-760,共5页Application Research of Computers

摘　　要：为进一步提高数据测试算法性能,提出一种基于控制流图支配树的测试数据灰度编码进化生成算法。利用三角分类程序示例构建数据测试的数据流控制流图,并利用其支配树关系构建测试数据的数据流分析,建立程序变量间的支配关系集。结合遗传算法并利用支配关系及分支距离构建测试数据生成的适应值函数,同时在应用遗传算法时,采用灰度编码方式取代二进制编码,简化编码更新过程。通过在测试程序样例中的实验对比显示,所提方法在平均数据生成数、平均收敛比率和数据生成数分布三个评价指标上得到明显提升。In order to improve the performance of data test algorithm,this paper proposed a gray encoding evolutionary algorithm with control flow dominating tree for test data generation. Firstly,it constructed the control flow chart of data test by using the triangle classification program,and then also constructed the data flow analysis of the test data by using the control tree,which established the relationship set between program variables. Secondly,combined with the genetic algorithm and using the dominance relation and branch distance,it constructed the fitness function of test data generation. Meanwhile,in the application of genetic algorithm,it used the gray encoding to replace the binary encoding,which could simplify the encoding update process. Finally,by comparing the experimental results in the test program,the proposed method significantly improves the three evaluation indicators of average number of data generation,the average convergence rate and the distribution of the number of data generated.

关 键 词：支配树 灰度编码 遗传算法 测试数据生成 支配关系 分支距离 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]