Research Of Urban Logistics Distribution Network Optimization Based On Constrained Clustering

Express Service is a kind of logistics business and can be tracked back to 50 years ago.With the rapid development of e-commerce,the development of express industry is also very fast.But this also leads to the poorly designed logistics distributon network,which has been a bottleneck restricting the development of e-commerce.Thus,it is extremely urgent to optimize the logistics distribution network key importance during designing a logistics distribution network.The siting of distribution stations is of key importance during designing or optimizing a logistics distribution network.This is because the distribution station has a direct relation with customers,and the evaluation of a customer to the express company is mainly based on the service quality of the distribution stations of this company,which can be significantly improved by the more reasonable siting of stations.Thus,in this paper we focus on the problem of the siting of distribution stations.This problem includes dealing with physical obstacles in real world,making the orders as close as possible to their respective stations and making the workload of each station as balanced as possible.To solve this problem,we use a dataset of the geographical coordinates of all orders of a certain express company per day in Shanghai,and propose an algorithm,which is a constrained clustering algorithm capable of handling physical obstacles and balance factor and outputting a set of clusters for decision-making.The main work and result are summarized as follows:First,a constrained clustering algorithm is proposed.It uses CLARANS as algorithm framework.This algorithm calculates the obstacle distance between two objects instead the Euclidean distance in consideration of physical obstacles and modifies the objective function to guide the clusters to having balanced workload.Second,the workload of a distribution station is defined.And the detailed deducing and accurate formula of the workload of a distribution are presented.Third,a system which can be used in the siting of distribution stations is designed and implemented.It implements the algorithm proposed by this paper,and provides the display and analysis of the result of clustering to help the analyst analysis the result.Besides,a comparison experiment is conducted on this system for examining the validity of the algorithm proposed in this paper.This paper introduces the current development of the express industry,describes the research background and significance.Then,the traditional clustering algorithms and the constrained clustering algorithms are summarized,and their disadvantages are analyzed.Meanwhile,the research ideas of this paper are proposed in the consideration of the problem in this paper.Then,this paper expounds the problem in detail,proposes a clustering algorithm which takes obstacles and balance factor into consideration,and defines the workload.Finally,a system is implemented based on this algorithm,and a comparison experiment is set up to proof the algorithm proposed in this paper works well.