Application Of Decomposition Multi-Objective Evolutionary Algorithm To Vehicle Routing Problem With Time Windows

In recent years,with the rapid development of my country's economy,the logistics field has entered an unprecedented development stage,and the logistics industry has attracted more and more attention from enterprises.However,my country's transportation and distribution costs account for about 52% of the total logistics costs.Improving transportation efficiency and reducing transportation costs are important ways to promote the stable development of the logistics industry.The vehicle routing problem is one of the hot research issues in the field of logistics distribution.Its research purpose is to rationally plan the vehicle distribution sequence in conjunction with the actual distribution situation so as to minimize the logistics cost.The vehicle routing problem is a typical multi-objective combination optimization problem.It is necessary to reasonably measure the control relationship between costs and determine the optimal solution through an algorithm.In this paper,based on the improvement of decomposition-based multi-objective evolutionary algorithm,it is used to solve the vehicle routing problem with time window.The main research contents of this article are as follows:(1)In this paper,the influence of fixed neighborhood on the traditional MOEA/D optimal decomposition multi-objective evolutionary algorithm is studied in depth,and a new improved neighborhood strategy is proposed.This strategy considers the degree of deviation of the subproblem from the central area and evolutionary algebra,and then proposes a new multi-objective decomposition evolutionary algorithm MOEA/D-INS,which updates the selected neighborhood and replaces the neighborhood,and dynamically adjusts the neighborhood size to balance the algorithm Convergence and diversity improve the uniformity of the solution set in the frontier of Pareto.Through the comparison experiment of algorithm performance effectiveness,this algorithm and other classic multi-objective algorithms are tested in the test function.Experimental results show that compared with several other algorithms,the overall quality of the solution set on the ZDT and DTLZ series test functions is significantly improved.(2)In this paper,the improved algorithm MOEA/D-INS is used to solve the multiobjective vehicle routing problem with time window.In terms of logistics cost,the vehicle distribution cost and the total distribution time are also considered to be minimized.Based on the location information of the distribution center and the distribution destination,the decomposition algorithm is used to decompose the multiobjective problem into a series of single-objective sub-problems,and optimize each sub-problem in parallel to obtain the vehicle path planning route.Compared with the path results obtained by the NSGA-II and MOEA/D algorithms,it can be seen that under the path planning solution obtained by the improved algorithm,the vehicle distribution cost and total delivery time are reduced,which proves that the improved algorithm proposed in this paper solves such problems.Effectiveness provides new ideas and methods for distribution center decision makers to provide better vehicle path planning solutions,reduce logistics costs,and solve multi-objective combination optimization problems with actual logistics background.