基于网络Voronoi图的大规模多仓库物流配送路径优化

摘要

Due to multiGconstraints and multiGobjectives,the optimization for large scale multiGdepot logistics routing problem is very difficult.A spatial heuristics algorithm is proposed based on the network Voronoi diagram.From the spatial perspective,two involved spatial issues in the multiGdepot logistics routing problem are service area partition and routing optimization.By using of depots’network Voronoi diagram,service area is coarsely partitioned and refined according to the goods storage in each depot.For the routing optimization,the local search space is limited within the spatial neighbors of customers.The proposed heuristics minimizes the used vehicles number and the total routes length.An experiment on several large scale logistics distribution instances in Shenzhen,China was implemented to validate the performance of the proposed heuristics algorithm.Results indicated that it provided high quality solution for large scale instances with 6400 customers in no more than 15 minutes.The proposed heuristics algorithm could be widely used in eGcommerce,express delivery,public utility in city to promote logistics efficiency.%由于存在多约束和多个优化目标，物流配送决策非常困难.本文针对城市多仓库物流配送问题，提出基于网络Voronoi 图的空间启发式优化方法.从空间角度将多仓库物流配送优化分解为区域分割和路径优化两个空间子问题.基于网络Voronoi 覆盖进行服务区域初始划分，顾及仓库容量差异，进行区域边界修正，并创建初始解.路径优化将局部搜索范围限定在网络 K近邻内，只搜索最有可能的空间邻域，迭代改进解的质量.该算法最小化路径数量和路径长度.利用深圳市的大规模多仓库物流配送问题测试算法性能.试验结果表明：本文方法能够在15 min 内求解6400个客户点的大规模物流配送问题，解的质量优于ArcGIS<10．8％，计算时间约为其21．2％.