The Research On Dispatch Models And Algorithms Of Express Delivery In Urban Logistics

With the rapid development of e-commerce and the change of consumption concept,online shopping has become the first choice for most people.At the same time,with the aggravation of urbanization,express delivery in the city has become the focus of enterprises and researchers.However,due to the popularity of shopping festivals such as "double 11" and "6.18",there is a lack of transportation capacity in the operation of express companies.Express delivery in the city is an urgent problem for enterprises to optimize and adjust to reduce operating costs.The problem of express delivery in the city studied in this paper is essentially a research category of VRP of city logistics.This paper focuses on the distribution from the distribution center to the service points,which is a vehicle scheduling problem with single source,multiple vehicle types and collection points that are used as secondary distribution centers.In this paper,based on the considerations of sufficient and insufficient transportation capacity of enterprises,we need to follow constraints of the time window requirements of service points,the load capacity constraints of vehicles and the traffic restriction policies of urban roads,and finally generate the vehicle scheduling plan to minimize transportation costs.Firstly,this paper studies the vehicle routing problem with sufficient capacity.For the scheduling plan with sufficient capacity,the mixed integer model is established.Firstly,a heuristic algorithm based on large neighborhood search is proposed to solve this problem quickly.Then,a branch and bound algorithm based on "node--vehicle" matching is designed.Finally,softmax regression,a machine learning algorithm,is added to guide the branch strategy in the branch and bound algorithm.Then,for the vehicle routing problem under the condition of insufficient transportation capacity,the method of intermediate delivery is introduced.According to the existing actual vehicles,the virtual capacity of the intermediate delivery is constructed.The mixed integer programming model is established,and the heuristic algorithm based on the priority of loading rate is designed to solve this problem.It is important to propose a branch and bound algorithm based on softmax regression to match the nodes and the vehicles appropriately.Finally,through the numerical experiments and case studies,the performance of the proposed algorithms are compared.The results show that the branch and bound algorithm performs well in solving these problems.Especially,the branch and bound algorithm based on machine learning shows strong adaptabilities,and can effectively improve the decisionmaking accuracy and speed in the real scheduling business.