Research On Dynamic Demand Based Vehicle Routing Problem With Time Windows

With the rapid development of smart logistics,the individual needs of customers are dynamic and uncertain,and the traditional vehicle routing model is no longer suitable for solving such problems.Therefore,the study of Dynamic Vehicle Routing Problem(DVRP)has great practical significance for the development of logistics industry.On the basis of domestic and foreign scholars on DVRP,this paper focuses on the dynamic demand based Vehicle Routing Problem(DDVRP),and uses an efficient intelligent optimization algorithm to solve it.The main research contents of this paper are as follows:(1)In order to satisfy logistics enterprises in the process of goods for distribution,there are three kinds of situations of new customer demand,customer cancellation service and customer change of delivery address,with the goal of minimizing the vehicle travel distance and the idea of combining pre-optimization and real-time optimization,a two-stage planning model of dynamic demand based vehicle routing problem with time windows(DDVRPTW)is established.The encoding method adopts integer encoding.In the pre-optimization stage,the improved genetic algorithm is used to obtain the pre optimized distribution route,the large-scale neighborhood search method is integrated in the mutation operation to improve the local optimization performance of the genetic algorithm,and a variety of operators are introduced to expand the search space of neighborhood solutions;In the real-time optimization stage,a periodic optimization strategy is adopted,and four kinds of neighborhood search operators are used to quickly adjust the path.Two different scale examples are designed for experiments.It is verified that the algorithm can plan the better route,and adjust the distribution route in time under the real-time constraints,which can provide theoretical guidance for suppliers to solve the dynamic demand based vehicle routing problem.(2)In response to the national call for energy conservation and emission reduction,green low-carbon logistics will become a new trend in the development of the modern logistics industry.The logistics systems of many enterprises have been constantly updated and improved,taking into account economic benefits and environmental benefits.In the past,the single-objective optimization that only considered the distribution cost of enterprises can no longer meet the distribution scenarios that consider multiple aspects in real life,and good customer satisfaction can improve the competitiveness of enterprises.Therefore,the optimization goal should be based on three perspectives of customers,logistics enterprises and the environment.Based on the DDVRPTW problem,this chapter studies the Multi-objective low carbon dynamic demand based Vehicle Routing Problem(MO-LCDDVRP)which simultaneously considers delivery costs,carbon emissions and customer satisfaction in the delivery process.The optimization goals mainly include vehicle fixed cost,transportation cost,penalty cost,carbon emission cost,and customer satisfaction;the dynamic demand elements involved mainly include new customer demand,original customer cancellation demand,and customer change of delivery address.Combined with the "pre-optimization and real-time optimization" strategy,a vehicle path planning model considering carbon emissions is established.In the pre-optimization stage,an improved simulated annealing algorithm is designed to solve the problem,and the greedy insertion method based on k-means is used to construct the initial solution of the simulated annealing algorithm.Generate the initial distribution path;in the real-time optimization stage,on the basis of the pre-optimized distribution plan,the idea of periodic optimization is used,combined with dynamic information,and various inter-path search operators and intra-path search operators are used to optimize the sub-paths,and then re-optimize the sub-paths.Plan delivery routes.Referring to the Solomon dataset and related literature,examples of different scales are designed and experiments are carried out to verify the effectiveness of the algorithm.