Research On Large-scale Logistics Regional Collaborative Distribution And Vehicle Routing Optimization

Logistics distribution,as the core link of the business transaction,not only guarantees the successful completion of business activities,but also is the key factor of enterprise profit and service quality.In addition,the advantages and disadvantages of urban logistics distribution have a direct impact on the traffic condition and environmental pollution of the city.In recent years,with the development of Internet technology and the rise of e-commerce,the traditional logistics distribution mode is developing towards a large-scale,multi-regional,intelligent direction.While achieving the purpose of distribution,how to effectively improve distribution efficiency,customer experience,cost control,and other indicators has become an urgent problem.However,the traditional logistics distribution method is difficult to solve large-scale urban logistics effectively.This paper mainly focuses on the large-scale logistics distribution of BBG chain supermarket,and the main research contents are as follows:1.Since the genetic algorithm is slow in evolution and easy to fall into the local optimal solution in the large-scale logistics distribution of single logistic center.This research established the mathematical model of the vehicle routing problem which considering vehicle capacity and maximum mileage and then designed a hybrid genetic algorithm.The method used greedy algorithm to initialize the population and hill-climbing the optimal individuals in each generation after undergoing selection,crossover,and mutation.And the experiment is validated on the dataset provided by BBG's commercial logistics management system.From the different experiments' results,it shows that the proposed method outperforms other methods in terms of distribution mileage.In the case of convergence of the algorithm,the method effectively speeds up the evolution of the algorithm while avoiding the phenomenon of “premature” in the population.2.Aiming at the problem of unreasonable vehicle routing planning and low loading rate in the large-scale multi-regional logistics distribution of a single logistics center,a regional collaborative distribution method based on road network topology is proposed.Firstly,based on the cone model in the spatial direction relation model,a relative direction relation model between regions based on fixed points is designed.And the distance relation between regions is added to construct the adjacency relation table of distribution area.Then,the distribution area is abstracted as the regional node on the vehicle distribution route,based on the area information in an actual distribution task and regional adjacency relation table,an initial route of the vehicle route distribution area is generated,and an inter-line area node is adjusted for the initial line which having the adjacent relationship,thereby forming a final route of the vehicle route distribution area.Finally,according to the overall arrangement of orders in the distribution area and single area S-GA,a regional collaborative distribution method is designed on the basis of vehicle distribution route.Compared with the regional independent distribution method,the regional collaborative distribution method can effectively improve the vehicle loading rate and reduce the distribution mileage.For the hybrid genetic algorithm,this paper improves the convergence rate of the algorithm and improves the quality of the solution.For the regional collaborative distribution method,this paper takes full consideration of the space position relation of the region,it breaks the situation that the resources in different regions are separated from each other in the traditional regional distribution,and realizes the cross-regional distribution of vehicles.