Research On Electric Vehicle Routing Problem For Flexible Delivery Scenarios

With the rapid development of e-commerce,more and more online shopping groups,the logistics industry has ushered in a booming period of development,and transportation as an important part of the logistics process is also facing opportunities and challenges.On the one hand,its costs account for 52.4% of the total social logistics costs.Therefore,in the face of increasingly diverse and complex distribution scenarios,further reducing transportation costs and improving transportation efficiency are difficult problems that logistics companies need to overcome.On the other hand,in urban distribution,Most logistics companies use traditional fuel vehicles for distribution.With the rapid development of the logistics industry,the number of logistics vehicles continues to increase and problems such as traffic congestion and environmental pollution caused by frequent urban distribution activities cannot be ignored.Pure electric vehicles are used as A new type of transportation tool,which has been widely used in various logistics scenarios with significant advantages of energy saving,low carbon and green,but it is affected by factors such as cruising range and inadequate recharging / swapping stations.Compared with traditional cars,they have higher use costs.In order to quickly respond to customers' distribution needs,formulate efficient distribution routes,and reduce the use cost of electric vehicles in the logistics distribution process,this paper studies the optimization of electric vehicle routing in three different distribution scenarios.First,in a static scenario,we study The Electric Vehicles Routing Problem with Soft Time window.We establish a mathematical model with the objective function of minimizing path cost,time window penalty cost,and vehicle use cost,and design mileage savings and improvements The tabu search algorithm solves the model.In addition,a new substation insertion strategy has been expanded,which not only considers the increased cost of the path caused by the insertion of the substation,the cost of changing the state of the nodes,and the penalty cost of creating an infeasible solution,but also the congestion caused by the substation Cost of waiting and driver anxiety cost.Finally,the feasibility and effectiveness of CW-TS algorithm are verified by comparing with genetic algorithm.In addition,a medium-scale example is given.Through sensitivity analysis,the results obtained under the two conditions of soft time window and mixed time window are compared and analyzed,as well as the effect of adjusting the parameters of the objective function and mileage anxiety threshold on the results.Secondly,in the dynamic scenario,we studied Dynamic electric vehicle routing problem based on staged solving strategy,and established a two-stage mathematical model.The first stage established an EVRP model for static customers,and the second stage established a dynamic EVRP model.To update the path.A formula for changing stations and dynamic customer insertion costs was constructed,and a hybrid heuristic algorithm for saving mileage plus an improved tabu search algorithm was designed to solve the model in the first stage,and the greedy algorithm was used to solve the model in the second stage to update the path.In order to verify the effectiveness and feasibility of the algorithm,a medium-scale example is given first.The results obtained by using CW and TS alone are compared with those obtained by using the CW-TS hybrid heuristic algorithm.The results show that CW-TS The algorithm obtains better results;secondly,by modifying the Solomon?100,C101 standard calculation example,the results show that the CW-TS algorithm can obtain a satisfactory solution within an ideal time.Finally,in a random scenario,we study Electric Vehicle Routing Problem with Stochastic Demands and Remedial Measure,establish a stochastic electric vehicle path optimization model with opportunity constraints,and design a mileage-saving algorithm plus an improved tabu search algorithm Hybrid heuristic algorithm to solve the model.In addition,a dynamic customer point insertion rule based on K-means clustering is provided,and a greedy algorithm is designed to solve it.Finally,the model and heuristic algorithm are tested in combination with cases and examples.The experimental results show that the CW-TS algorithm can find a better solution in less time than the TS algorithm.