计算机应用 ›› 2012, Vol. 32 ›› Issue (04): 1165-1167.DOI: 10.3724/SP.J.1087.2012.01165

• 典型应用 • 上一篇    下一篇

基于K-均值聚类粒子群优化算法的组合测试数据生成

潘烁,王曙燕,孙家泽   

  1. 西安邮电学院 计算机学院, 西安 710061
  • 收稿日期:2011-09-21 修回日期:2011-11-17 发布日期:2012-04-20 出版日期:2012-04-01
  • 通讯作者: 潘烁
  • 作者简介:潘烁(1987-),女,河北鹿泉人,硕士研究生,主要研究方向:软件测试;
    王曙燕(1964-),女,河南南阳人,教授,博士,CCF会员,主要研究方向:软件测试、数据挖掘;
    孙家泽(1980-),男,河南南阳人,讲师,博士研究生,主要研究方向:软件测试、智能信息处理。
  • 基金资助:
    国家自然科学基金资助项目

Test data generation based on K-means clustering and particle swarm optimization

PAN Shuo,WANG Shu-yan,SUN Jia-ze   

  1. School of Computer Science and Technology,Xian University of Posts and Telecommunications,Xi’an Shaanxi 710061, China
  • Received:2011-09-21 Revised:2011-11-17 Online:2012-04-20 Published:2012-04-01
  • Contact: PAN Shuo

摘要: 在解决组合测试中的测试数据集生成问题时,粒子群优化算法(PSO)在待测数据量增加达到一定程度以后,出现迭代次数增加、收敛速度减慢的缺点。针对该问题,提出了一种应用于组合测试数据集生成问题的基于K-均值聚类的粒子群优化算法。通过对测试数据集合进行聚类分区域,增强测试数据集的多态性,从而对粒子群优化算法进行改进,增加各个区域内粒子之间的影响力。典型案例实验表明该方法在保证覆盖度的情况下具有一定的优势和特点。

关键词: 组合测试, 粒子群优化算法, K-均值聚类算法, 测试数据

Abstract: To solve the problem of the test data set generation in combinatorial test, if the software under test has a great many factors and values, the traditional Particle Swarm Optimization (PSO)will have large iteration times and slow convergence velocity. A test data set generation method based on K-means clustering algorithm and PSO has been proposed. The polymorphism of the test data set has been enhanced, though clustering and partitioning the test data set. And it makes PSO has been improved. The compact between the particles in each area has been promoted. Several typical cases show that this method has some merits while ensuring the coverage.

Key words: combinatorial test, Particle Swarm Optimization (PSO) algorithm, K-means clustering algorithm, test data

中图分类号: