Research On Location Selection Of Logistics Distribution Center Based On Bayesian Network

With the continuous development of e-commerce and express logistics,users'requirements for personalized services such as timely delivery,timing,and door-to-door are constantly improving.At the same time,higher requirements are imposed on logistics timeliness and transportation costs.As a key node in the logistics network,the distribution center directly affects the timeliness and costs.This article elaborates the relevance theory of the distribution center and the related theory of Bayesian network.According to the characteristics of the distribution center,the location method of the logistics distribution center based on Bayesian network is proposed,including four parts.The first part is the Bayesian network modeling stage.Based on the Bayesian network,the paper analyzes the relevant characters of the distribution center in the country and models the Bayesian network between the characters of the distribution center.The second part is to obtain candidate locations,select candidate locations based on POI constraints and existing warehouse resources.The third part is to filter the candidate locations,apply the Bayesian network model library,calculate the posterior probability of the candidate location-related attributes,and delete and classify the nodes.The fourth part is to design the layout of the logistics network,select different number of candidate locations from each set,meet the coverage constraints,combine the order data,calculate the system objective function,and count the frequency of the nodes in the multiple systems that meet the conditions.The descending order of the frequency of the nodes is selected,and then the nodes are selected to satisfy the coverage constraint,and the objective function is calculated again,and the system under the optimal objective function value is taken as the final layout.Through the above four parts,a method of selecting a distribution center in a given area is obtained.The algorithm proposed in this paper is simulated and compared with other algorithms,and the effectiveness of the algorithm is verified.