求解多目标带时间窗VRP的文化狼群算法  

Cultural wolf pack algorithm for solving multi-objective VRP with time window

作　　者：李小川 刘媛华[1] 王影歌 Li Xiaochuan;Liu Yuanhua;Wang Yingge(Business School,University of Shanghai for Science&Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学管理学院,上海200093

出　　处：《计算机应用研究》2020年第4期1025-1029,共5页Application Research of Computers

基　　金：国家自然科学基金资助项目(11505114);国家教育部人文社会科学研究一般项目(12YJC630127)。

摘　　要：针对以最小化总距离和车辆数为目标的带时间窗车辆路径问题的求解,提出一种基于文化基因的狼群算法。根据客户近邻度矩阵构建初始狼群,提高狼群的优良性。结合问题特征与狼群算法的寻优思想,重新定义了群体的游走和突袭行为。提出头狼变异策略,增加了最优解的邻域搜索范围。受文化算法启发,将当前最优解的有利信息作为信仰空间的知识来指导狼群进化方向,提高搜索效率。仿真实验及与其他算法对比表明,文化狼群算法求解车辆数更少、总距离更短,且具有更好的稳定性。To solve the vehicle routing problem with time window for the shortest length of vehicle travel and the minimum number of the using vehicles,this paper proposed a cultural wolf pack algorithm.To improve the quality of the wolf pack,this method constructed the initial wolf pack according to the customer’s nearest neighbor matrix.Considering the features of this problem and the wolf pack algorithm,it redefined the intelligent behaviors of artificial wolves.The mutation of leader wolf increased the searching range of the optimal solution.According to the favorable information of the current optimal solution,it constructed the knowledge of the belief space to guide the evolution direction of the wolf pack and improve the search efficiency.Simulation results show that the cultural wolf pack algorithm can get less number of vehicles,shorter total distance and better stability than other algorithms.

关 键 词：车辆路径问题 时间窗 多目标 文化狼群算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]