Optimal Location Of Logistics Distribution Center Based On Improved K-means Clustering Algorithm

Online shopping has become indispensable in our lives,logistics distribution is also a very key aspect in it.Because of the great impact on distribution efficiency and operation cost,the topic about optimal location of logistics distribution has always been the hot spot and difficulty of research.K-means clustering algorithm is a very effective optimization location method,but it also has many problems.In order to improve the distribution efficiency and reduce the operation cost,aiming at the shortcomings of the traditional K-means clustering location algorithm,such as "subjective setting of k-value selection","distribution time is not considered in performance optimization" and "the address determined by the algorithm can not build a logistics center in reality",taking "optimal location selected of logistics distribution center address for Shaanxi ZTO Express Co.,Ltd." as the hypothetical proposition,a network function structure diagram based on the four-level logistics distribution center is designed,and the data set is established by crawling the data of the basic level of CAINIAO Post station in Xi'an.To get optimal distribution time,the problems of number determination and positioning of the second level distribution center(Urban Logistics Center)and the optimal location of the third level distribution center(regional distribution center)are studied.An improved k-means regional distribution center optimal location algorithm based on distribution time optimization and a clustering location algorithm of realizable urban logistics center are proposed.In order to solve the problem that traditional K-means algorithm does not consider the optimization of distribution time,an improved k-means optimal location algorithm for regional distribution center based on distribution time optimization is proposed.The data set is constructed by using the data about the basic level CAINIAO Post station.Firstly,K clusters are obtained by using of the traditional K-means clustering algorithm.Then,points that can lead to the minimum distribution time sum to other points in every cluster are determined according to the principle of TSOTSJ(the distribution time in cluster),and these points are used as the regional distribution centers.The simulation example takes the existing 1052 CAINIAO Post stations as the data set,and the TSOTJ obtained by the algorithm proposed in this paper is compared with the traditional K-means clustering algorithm.The results show that the proposed algorithm can save about 66437 s for delivery time,about 10.72% shorter than K-means clustering algorithm.The distribution center determined by the traditional K-means clustering algorithm perhaps can't be built in real life because of the poor construction environment.To solve this problem,a clustering location algorithm with good construction condition for urban logistics center is proposed.Firstly,some important attributes affecting the number of urban logistics centers are mined by training logistics distribution centers data in other cities.Taking these attributes of Xi'an as the test set,the number of the second-level distribution centers is predicted.Then,Xi'an is devided into K parts.In every part,several places suitable to build the second-level distributions center are selected for alternative construction addresses of the second-level distribution center.Selecting the best K address which have the minimum total distribution time from the alternative address to the thirdlevel distribution centers,they are the K location of the second-level logistics distribution center.The sum of total distribution time of all clusters proposed in this paper is compared with the clustering location algorithm of urban logistics center based on distribution distance by some simulation examples.The results show that the urban logistics center location algorithm proposed in this paper can save about 10476 seconds to complete one distribution to 30 regional distribution centers,about 20.18% shorter than the urban logistics center clustering location algorithm based on distribution distance.Moreover,the determined urban logistics center addresses are in the alternative construction address set,It can be realized in reality.