Research On Multi-agent Cold Chain Logistics Vehicle Routing Problem

With the improvement of living standard,people's demand for cold chain logistics is getting higher and higher.However,cold-chain logistics still has not formed a complete service system.The two problems of high distribution cost and distribution efficiency restrict the development of whole cold chain logistics.Reducing the cost of cold chain logistics and improving distribution efficiency has become a hot topic of social concern.The optimization of distribution path is the key point to save logistics transportation cost,so the study of path optimization is very important.At the same time,in many path optimization algorithms,genetic algorithm has good robustness,extensibility and intelligence in solving optimization problems.Therefore,in this paper,the system structure of cold chain logistics transportation scheduling system is constructed by using multi-agent system.In addition,this paper establishes a reasonable model to optimize the path and reduce the distribution cost.At the same time,this paper solves the problems of slow convergence and precocity of traditional genetic algorithm by improving the genetic algorithm.The main contents of this paper are as follows:Firstly,based on the optimization of the urban cold chain logistics distribution path,this paper analyzes the related knowledge of the cold chain logistics and the vehicle path problem area with time window.At the same time,the research results and research methods of vehicle routing problem(VRP)are summarized.Secondly,this paper analyzes the cost of urban cold chain logistics distribution,and introduces the theory and modeling method of multi-agent technology.Based on the basic model of VRP,this paper constructs a transport scheduling model of cold chain logistics enterprise based on multi-agent technology.Thirdly,aiming at the problem of slow convergence speed and early maturity of traditional genetic algorithm,a new improved genetic algorithm is proposed.First,the greedy algorithm is introduced to initialize the population,which makes the population better.Then,the clustering method of minimum spanning tree is adopted to improve the crossover operation and increase the diversity of the population.Finally,the convergence rate of the algorithm is accelerated by the combination of the Symbiotic Organisms Search algorithm(SOS).The improved genetic algorithm converges faster and avoids the local optimal problem.Finally,an example is given to verify the validity of the proposed model and the improved genetic algorithm.This is a step-by-step approach to the optimization of the chain of logistics distribution.And the conclusion of this paper has some practical significance for improving the distribution efficiency and logistics service level.