基于智能非信息素蜂群优化的软件测试研究  

Intelligent non-pheromone swarm optimization based software test suite

作　　者：李云玮[1] 孙忱[2] 范玉顺[3] 

机构地区：[1]北京电子科技职业学院电信工程学院,北京100015 [2]中国科学技术大学信息科学技术学院自动化系,合肥530001 [3]清华大学自动化系,北京100083

出　　处：《计算机应用研究》2014年第8期2399-2402,共4页Application Research of Computers

基　　金：北京市教委资助性项目(20130834)

摘　　要：针对测试套件优化问题,提出了一种基于非信息素的人工蜂群优化方法。该方法将每个测试实例都看做优化问题的一个可能解,并引入幸福值用于评价测试实例的吻合程度。通过将三组蜂群分别扩展为搜索代理、选择代理和优化代理,可以从大量的测试实例中选出有效的测试实例。利用这些代理的并行特性,并使用路径覆盖范围作为测试充分性准则,提高了测试实例在每次迭代过程中的路径覆盖范围,加快了解的产生速度,从而提高了该方法的运行速度和效率。仿真结果比较了该方法与蚁群优化算法的性能,证明了该方法的收敛速度和优化质量均优于蚁群算法。This paper proposed a non-pheromone based artificial bee colony optimization（ ABC） for test suite optimization.The approach considered each test case as a possible solution in the optimization problem and introduced happiness value to evaluate the quality of fitness of the test cases. Through extending three groups of bees to three agents namely search agents,selector agents and optimizer agents,efficient test cases could be selected among near infinite number of test cases. Taking advantage of the parallel behavior of these agents,and using path coverage as the test adequacy criterion,it improved the path coverage of the test cases in each iteration. It accelerated the solution generation,and therefore improved the running speed and efficiency of the proposed approach. Compared the performance of the proposed approach with that of the ant colony optimization（ ACO）,simulation results verify that both the convergence rate and optimization quality of the proposed approach are better than ACO based optimization.

关 键 词：软件测试 测试优化 人工蜂群优化 测试充分性准则 蚁群优化 

分 类 号：TP311.53[自动化与计算机技术—计算机软件与理论]