基于视觉修正的改进最大最小蚂蚁系统求解TSP  

IMPROVED MAX-MIN ANT SYSTEM BASED ON VISION CORRECTION FOR SOLVING TSP

作　　者：李克文[1] 徐延辉 张震涛 席英杰 Li Kewen;Xu Yanhui;Zhang Zhentao;Xi Yingjie(College of Computer Science and Technology,China University of Petroleum,Qingdao 266580,Shandong,China)

机构地区：[1]中国石油大学(华东),山东青岛266580

出　　处：《计算机应用与软件》2023年第10期279-284,328,共7页Computer Applications and Software

基　　金：国家自然科学基金重大项目(51991361);国家科技重大专项(2016ZX05021-002)。

摘　　要：蚁群算法的研究与发展为解决优化问题提供了良好的方案,但是依然存在收敛速度慢和容易陷入局部最优的问题。为此,提出一种基于视觉修正的VC-MMAS算法。该算法以启发式信息为依据初始化信息素浓度,从而加速初期的算法收敛速度;引入“天文蚁”对历史最优路径进行检查和修正,同时使用“双优”策略更新信息素,帮助算法跳出局部最优。在旅行商问题(TSP)上的实验对比表明,该算法与其他算法相比具有更好的全局寻优能力和稳定性。The research and development of ant colony algorithm provides an efficient solution for the optimization problem.However,the ant colony algorithm also has defects such as slow convergence speed and falling easily into local optimum.This paper proposes a max-min ant system algorithm based on vision correction(VC-MMAS).The proposed algorithm initialized pheromone concentration based on heuristic information,which raised the initial convergence speed of the algorithm.At the same time,the algorithm used "astronomical ant" to check and modify the historical optimal path,and the "double optimization" strategy to update pheromone,making the algorithm exceed the local optimum.The experimental results of traveling salesman problem(TSP) show that the algorithm has better global optimization ability and stability than other algorithms.

关 键 词：最大最小蚂蚁系统 启发式信息 信息素 局部最优 旅行商问题 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术] TP18