Research On Dynamic Vehicle Routing Problem With Integration Of Pickup And Delivery

With the rapid global economic growth and the rapid development of the Internet,the logistics industry is facing enormous challenges.Transportation costs and distribution efficiency directly affect the overall development of the logistics industry.In order to further improve the level of customer service and satisfaction,logistics distribution mode and routing optimization have become the focus of research in all aspects.However,with the continuous expansion of customer scale and the gradual increase of personalized demand,it is difficult for the basic Vehicle Routing Problem(VRP)model to meet the distribution requirements of logistics enterprises.Therefore,this article selects the uncertain needs of customers as the research object,relying on the integration of pickup and delivery,in order to better improve the efficiency of delivery and meet the various needs of customers,it provides reference for logistics enterprises in solving the dynamic vehicle routing problem with integration of pickup and delivery.This article mainly has the following work:(1)Summarizing and analyzing the current domestic and foreign research status of dynamic vehicle routing problem and vehicle routing problem with integration of pickup and delivery,and detailed description of the Dynamic Vehicle Routing Problem with Integration of Pickup and Delivery(DVRPIPD).The impact of dynamic events on the original delivery route is analyzed,and a dynamic event conversion strategy is proposed,which greatly reduces the complexity of solving dynamic events when they appear.(2)A DVRPIPD problem model with hard time windows were constructed.The construction of the initial delivery route can be effectively completed by creating a real-time monitoring function for loading.The real-time service recording function can quickly make route adjustment to unserved customers.(3)A two-stage solution algorithm is designed for the DVRPIPD.In the initial optimization stage,combined with the characteristics of the global breadth search of the brain storm optimization algorithm and the local depth search of the variable neighborhood search algorithm,a hybrid variable neighborhood search brain storm optimization algorithm is designed.In the early stage of the algorithm,in order to quickly obtain better individuals,the ant colony optimization algorithm was used to construct the initial population.By using the breadth search capability of the brain storm optimization algorithm to select individuals to be evolved,and then using the Exchange,Relocate and 2-opt within the path in the variable neighborhood search algorithm,and the Swap,Shift and 2-opt* between the paths to adjust the path,finally can quickly find the optimal solution.In the dynamic optimization stage,with the help of the greedy algorithm to quickly find the best solution in the current situation,coupled with the advantage of the dynamic event conversion strategy to simplify the complexity of the problem,a greedy insertion optimization algorithm is designed to solve the real-time path optimization problem.(4)Through the comparison of the test results of small and medium-scale and large-scale Solomon standard data,the effectiveness and stability of the hybrid variable neighborhood brain storm optimization algorithm are verified.In order to adapt to the DVRPIPD,verify the model and solution algorithm in this article,and modify the Solomon data.Then the sensitivity analysis of the parameters of the DVRPIPD is carried out.Finally,a large-scale DVRPIPD is used to specifically analyze the solving ideas of the algorithm in this article,and it is verified that the algorithm in this article can effectively solve the DVRPIPD.