计算机科学与探索 ›› 2014, Vol. 8 ›› Issue (8): 1002-1008.DOI: 10.3778/j.issn.1673-9418.1307009

• 人工智能与模式识别 • 上一篇    下一篇

改进的重叠蚁群优化算法

彭  岳1,2,王  俊1,2,谢斌福1,2,张月峰1,2,王崇骏1,2+   

  1. 1. 南京大学 计算机科学与技术系,南京 210023
    2. 南京大学 软件新技术国家重点实验室,南京 210023
  • 出版日期:2014-08-01 发布日期:2014-08-07

Improved Overlap Ant Colony Optimization Algorithm

PENG Yue1,2, WANG Jun1,2, XIE Binfu1,2, ZHANG Yuefeng1,2, WANG Chongjun1,2+   

  1. 1. Department of Computer Science and Technology, Nanjing University, Nanjing 210023, China
    2. State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing 210023, China
  • Online:2014-08-01 Published:2014-08-07

摘要: 针对蚁群优化算法在进行全局最优解搜索时容易陷入局部最优解和收敛速度缓慢等缺陷,提出了一种有效求解全局最优解搜索问题的重叠蚁群优化算法。该算法通过设置多个重叠的蚁群系统,并对每一个蚁群初始化不同的参数,之后在蚁群之间进行信息素的动态学习,增强了不同蚁群对最优解的开采能力,避免了算法出现早熟现象。仿真实验结果表明,重叠蚁群优化算法在避免陷入局部最优解方面具有良好的效果,是一种提高蚁群算法性能的有效的改进算法。

关键词: 蚁群优化(ACO), 局部最优解, 重叠蚁群优化, 动态学习

Abstract: Considering that the ant colony optimization (ACO) algorithm in solving global optimal solution search problems has the defects such as slow convergence and prone to local optimal solution phenomenon, this paper provides an overlap ant colony optimization (OACO) algorithm to solve global optimal solution search problem. By setting multiple overlapping ant colony, initializing different parameters of every ant colony, and learning dynamically between ant colony pheromones, OACO algorithm reinforces the exploitation of ant colony optimization algorithm and avoids premature phenomenon. The experimental results show that OACO algorithm achieves good results in avoiding falling into local optimal solution, and is an efficient and effective improved ant colony optimization algorithm.

Key words: ant colony optimization (ACO), local optimal solution, overlap ant colony optimization, dynamic learning