基于多态蚁群算法的测试用例自动生成  

Automated software test-data generationbased on polymorphic ant colonies algorithm

作　　者：陈明师[1] 刘晓洁[1] 李涛[1] 

机构地区：[1]四川大学计算机学院,成都610065

出　　处：《计算机应用研究》2009年第6期2347-2348,共2页Application Research of Computers

基　　金：国家自然科学基金资助项目(60573130;60502011);国家教育部新世纪优秀人才计划资助项目(NCET-04-0870);国家"863"计划资助项目(2006AA01Z435);国家教育部博士点基金资助项目(20070610032)

摘　　要：提出了一种基于多态蚁群算法的测试数据自动生成方法。该方法使用二进制编码将输入数据转换为位串;然后在蚁群算法的基础上将蚁群分为三类,据其信息素的不同采用不同的移动准则,重点对侦察蚁和搜索蚁进行功能分析。将局部搜索与全局搜索结合起来,结合路径的相似度,缩小搜索空间;根据适应度函数确定最好路径,既解决局部最优化问题,又提高收敛效率。与基本蚁群算法对比,其结果显示该方法效率优于基本蚁群算法。This paper drawled.a polymorphic ant colonies algorithm method to generate test-data automatically. This method transferred the data input to bit string in binary coding approach. Based on the ant colonies algorithm, divided the ants into three categories. Different pheromones with different criteria to move, focused the analysis of exploring ant and searching ant. combined local search and global search to explore the optimal solution. With the similarity of the route, it could narrow the search space and increase the convergence rate; with the fitness function, it could find the optimal routes. This method could solve the local optimal problems and improve the efficiency of convergence. Comparing and studying the ant colony algorithm, show that the approach is feasible and superior to the previous in efficiency.

关 键 词：多态蚁群算法 测试数据 相似度 适应度 信息素 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]