计算机工程与应用 ›› 2009, Vol. 45 ›› Issue (27): 11-14.DOI: 10.3778/j.issn.1002-8331.2009.27.004

• 博士论坛 • 上一篇    下一篇

一种改进的基于隶属云模型的蚁群算法

张煜东,吴乐南,韦 耿   

  1. 东南大学 信息科学与工程学院,南京 210096
  • 收稿日期:2009-06-11 修回日期:2009-07-11 出版日期:2009-09-21 发布日期:2009-09-21
  • 通讯作者: 张煜东

Improved ant colony algorithm based on membership cloud models

ZHANG Yu-dong,WU Le-nan,WEI Geng   

  1. School of Information Science & Engineering,Southeast University,Nanjing 210096,China
  • Received:2009-06-11 Revised:2009-07-11 Online:2009-09-21 Published:2009-09-21
  • Contact: ZHANG Yu-dong

摘要: 为了解决传统蚁群算法的收敛速度慢和易陷入局部最优等缺陷,做出如下改进:首先采用云模型来自适应控制蚂蚁的随机性;其次缩小了后继城市的搜索范围;最后引入2-opt局部搜索策略。对城市规模从50到高达1 000的TSP问题进行仿真,并与先前提出的改进蚁群算法进行对比,结果表明,该算法不仅偏离率更小,而且运行时间短。随着城市规模的增大,优势更明显。

关键词: 隶属云, 蚁群算法, 旅行商问题

Abstract: To overcome the slow convergence and local extrema of ant colony algorithm(ACA),improvements are proposed in this article.Firstly,control the randomness of the ants via cloud model;secondly,reduce the search region of subsequent cities.Finally,introduce the 2-opt local search strategy.Simulations are implemented on various TSP with cities from 50 to 1 000,and this proposed algorithm is compared with old algorithm.Results demonstrate that the gap of this proposed algorithm is lower and the computation time is shorter than old methods,the advantages becoming notable with the number of cities becoming larger.

Key words: membership cloud, ant colony algorithm, traveling salesmen problem

中图分类号: