一种面向对象测试用例自动生成的混合算法  

Hybrid algorithm of automatically generating of test data for object-oriented program

作　　者：胡岳峰[1] 高建华[1] 

机构地区：[1]上海师范大学数理信息学院计算机科学与工程系,上海200234

出　　处：《计算机应用研究》2008年第3期786-788,802,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(60673067);上海市科委自然科学基金资助项目(04ZR14105)

摘　　要：遗传算法(GA)在测试用例生成方面是一种实用的算法,但是其自身也存在的局限性,如过早收敛、优化效率低等问题。通过引入粒子群算法(PSO),使每一个测试用例在局部区域中再次寻找最优值,以此改进整体算法搜索最佳值的能力,避免过早收敛、优化效率低的问题。与此同时,针对面向对象测试的特点,如封装性等,将混合算法进行适当的改进,满足在不同环境中重复使用类的要求。Genetic algorithm is an effective algorithm in automatic generation of test data. But it also has the deficiency such as ahead convergence and short of climbing ability, etc. This paper improved the ability of searching for the most suitable value and avoided the problem of ahead convergence and short of climbing ability by introducing the particle swami optimization （PSO）. Meanwhile this paper also improved the hybrid algorithm in order to meet the needs of using the object-oriented program in different conditions.

关 键 词：软件测试 测试数据 遗传算法 粒子群优化 

分 类 号：TP311.56[自动化与计算机技术—计算机软件与理论]