基于FCM-PSO-FWA算法的多个菜鸟驿站选址方法

Multi-location Selection Method for Cainiao Stations Based on FCM-PSO-FWA Algorithm

聚焦社区内多个菜鸟驿站选址问题,通过融合多种机器学习方法,设计出能有效解决多个菜鸟驿站选址问题的算法,并运用算例仿真验证所提算法的有效性。首先根据取件方式建立基于0-1整数规划的菜鸟驿站选址模型;接着运用模糊C均值(Fuzzy C-means,FCM)聚类算法将一个多菜鸟驿站选址问题划分成多个单一菜鸟驿站选址问题;然后将烟花算法(Fireworks Algorithm,FWA)融入粒子群算法(Particle Swarm Optimization,PSO),解决传统PSO算法易陷入局部最优解的问题;最后将所提FCM-PSO-FWA算法应用于随机生成不同规模需求点的菜鸟驿站选址和江南大学菜鸟驿站选址中。算例结果表明:在任意数量的需求点条件下,FCM-PSO-FWA算法能有效地解决多个菜鸟驿站的选址问题,并且不会陷入局部最优解,从而验证了该算法的可行性和有效性。

With the popularization of e-commerce and the rapid growth of consumer demand for express delivery services,the location allocation of the Cainiao station,the facility for a new mode of express delivery service,is crucial to improving the efficiency and user experience of express delivery services.Most current studies focus on the location allocation of logistics distribution centers and,of the relatively small body of re⁃searches on the location allocation of the Cainiao station,the majority consider the location allocation of a single logistics distribution center,despite the fact that realistically,communities such as university campuses and large residential areas often need to establish multiple Cainiao stations.In this paper,we design an algorithm to solve the location allocation problem of multiple Cainiao stations in a community by integrating multiple machine learning methods,and verify the effectiveness of the pro⁃posed algorithm through a simulation example.First,we establish the Cainiao station location allocation model according to the parcel pickup method based on 0-1 integer programming,and use the particle swarm optimization algorithm(PSO)to find the optimal location of the Cainiao station.However,PSO can only solve the optimal location problem of a single Cainiao station through iteration,which is not viable in the case of multiple Cainiao stations.In light of this,we employ the fuzzy C-means(FCM)clustering meth⁃od to conglomerate all demand points according to their geographical coordinates and obtain the cluster cen⁃ter of each type of demand points.Then such cluster center is regarded as the initial location of the Cainiao station for each type of demand point and in this way,we divide the location allocation problem of the multi⁃ple Cainiao stations in a community into the location allocation problem of a single Cainiao station in multi⁃ple sub-communities.Next,we incorporate the Fireworks Algorithm(FWA)into the Particle Swarm Opti⁃mization(PSO)to deal with the locality tendency of the traditional PSO algorithm and finally apply the pro⁃posed FCM-PSO-FWA algorithm to the Cainiao station locatioin allocation problem with randomly gener⁃ated demand points of different sizes and the Cainiao station location allocation problem of Jiangnan Univer⁃sity.The study shows that given undefined number of demand points,the FCM-PSO-FWA algorithm can ef⁃fectively solve the location allocation problem of multiple Cainiao stations and will not fall into local conver⁃gence,thus verifying the feasibility and effectiveness of the algorithm.