Research On Cold Chan Logistics Terminal Joint Distribution Based On Complex Network And Machine Learning

In recent years,with the improvement of living standards,people's consumption concept has changed.On the basis of solving the problem of food and clothing,people are increasingly pursuing the quality and nutrition of food.The upgrading of consumption directly promotes the development of cold chain logistics industry in China.However,due to the late start of cold chain logistics in China,the development level is relatively backward,the overall operation efficiency is low,and the cost remains high for a long time.Especially in the terminal distribution link,the quality of service is often not guaranteed,which is one of the main causes of cold chain "broken chain"and goods damage.Faced with this situation,this paper introduces the idea of co-distribution into cold chain logistics,designs and implements a set of co-distribution schemes for cold chain logistics based on complex network and machine learning,aiming at improving the efficiency of cold chain logistics,reducing the cost of distribution and ensuring the quality of service.Firstly,this paper summarizes the research status of cold chain logistics and joint distribution at home and abroad.Secondly,on the basis of defining the related concepts and contents of cold chain logistics,the necessity of end-to-end joint distribution is analyzed and described.Then,according to the actual situation of the end of the city,a complete end-to-end joint distribution scheme is designed from three aspects:the regional division of joint distribution,the mode selection of distribution service and the income distribution of cooperation.Taking some areas of Beijing "city deputy center"as an example,the feasibility of the scheme is verified.In the first step of the scheme,a community network model is established through specific transformation rules,and the analytic hierarchy process is introduced in the process of building the model.The difficulty of inter-cell distribution is evaluated comprehensively,and its value is taken as the weight of each side in the network model.On this basis,the Louvain algorithm is used to solve the community discovery problem of the network model,and the final result of the division of the terminal distribution area is obtained.In the second step,Adaboost classification algorithm is used to select the service mode of each distribution area.By collecting relevant feature data and training machine learning model,the algorithm classification is realized and the corresponding service modes of each region are obtained.Thirdly,through the improved Shapley value method,the income distribution problem of joint distribution alliance is studied.This paper proposes a revenue allocation scheme based on weighted Shapley value method,which fully considers the importance of each enterprise in cooperation.Finally,through the calculation and analysis of case data,it is found that the results of the scheme are consistent with the actual situation,which has strong feasibility,and provides a new idea and direction for the research of cold chain logistics end-to-end joint distribution.