基于混沌退火粒子群优化算法的路径测试数据生成  

Path Test Data Generation Based on Chaos Anneal Particle Swarm Optimization Algorithm

作　　者：陈策[1,2] 赵春霞[1] 

机构地区：[1]南京理工大学计算机科学与技术学院 [2]63961部队

出　　处：《南京理工大学学报》2011年第3期376-381,共6页Journal of Nanjing University of Science and Technology

摘　　要：为实现指定路径的软件测试数据自动生成,提出了一种基于粒子群优化(PSO)算法的演化测试方法。利用分支函数插装和强制路径执行策略,得到用于优化搜索的路径适应值。通过引入混沌搜索、模拟退火和早熟收敛判断机制,克服了标准PSO算法易陷入局部最优而无法找到测试数据的缺陷。三角形判断程序的测试数据自动生成实验表明:在最大迭代次数Tmax为500时,混沌退火粒子群优化(CAPSO)算法的命中概率为99%,标准PSO的命中概率为95%;在Tmax为2 000时,CAPSO算法的命中概率为100%,标准PSO算法的命中概率为95%左右;继续增大Tmax不能使标准PSO算法的命中概率提高,而CAPSO算法总能摆脱局部极值找到满足要求的测试数据。A kind of evolutionary test method based on the particle swarm optimization（PSO）algorithm is proposed for the automatic generation of appointed path software test data.Path adaptive value for optimization searching is calculated using code insertion of branch functions and control execution strategy of program path.The shortcoming that the standard PSO algorithm is easy to fall into local optima and can ＇ t find the test data is overcome by introducing chaos search,simulated annealing and prematurity convergence judgment mechanism.The experiments of the automatic generation of test data on a triangle judgment program show：when the biggest iterative time Tmax is 500,the hit probability of chaos anneal particle swarm optimization（CAPSO）algorithm and standard PSO algorithm is 99% and 95%;when Tmax is 2000,the hit probability of CAPSO algorithm and standard PSO algorithm is 100% and 95%;the increase of Tmax can ＇ t improve the hit probability of the standard PSO algorithm,but the CAPSO algorithm can shake off the local extremum and find the test data satisfying the request.

关 键 词：粒子群优化 模拟退火 混沌搜索 早熟收敛判断 软件测试 路径测试 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置] TP31[自动化与计算机技术—控制科学与工程]