应用于车辆路径问题的多蚁群算法研究.pdfVIP

th Proceedings of the 25 Chinese Control Conference 7-11 August, 2006, Harbin, Heilongjiang 应用于车辆路径问题的多蚁群算法 1 1 1 2 陈宝文 ，宋申民 ，陈兴林 ，单志众 1. 哈尔滨工业大学 航天学院, 哈尔滨 150001 E-mail: chenbaowen@ 2. 杭州贝尔通讯系统有限公司, 杭州 310007 E-mail: shanzhizhong@263.net 摘 要: 针对传统蚁群算法容易出现早熟和停滞现象的缺陷, 提出了一个基于MMAS的改进蚁群算法. 在蚁群算 法的收敛性证明和对参数的实验分析基础上, 采用多蚁群协作, 修改信息素更新规则和根据收敛要求动态调整 主要参数等对蚁群算法进行改进, 应用该方法解决带时间窗口的车辆路径问题. 实验结果表明此方法有更好的 全局搜索最优解的能力并能有效的减少收敛时间. 关键词: 蚁群算法, 车辆路径问题, 启发式算法 A Multi-Ant Colony System for Vehicle Routing Problems 1 1 1 2 Chen Baowen , Song Shen-min , Chen Xinglin , Shan Zhizhong 1. School of Astronautics, Harbin Institute of Technology, Harbin 150001, China E-mail: chenbaowen@ 2. Hang Zhou Bell Telecommunication System Co. Ltd. Hangzhou 310007, China E-mail: shanzhizhong@263.net Abstract: In order to accelerate the convergence rate of the ant colony algorithm and avoid the stagnation behavior, a novel ant colony system developed from MAX-MIN ant system is present. The methods including Multi-ant colony, improving state transition rules, introducing parameter mutation method according the convergent requirement are used in the algorithm. It is applied to the vehicle routing problem with time window. Experimental results show that algorithm has much higher capacity of global op