The Research On The Optimization Of Distribution Routes For Fresh Agricultural Products

With the rapid development of online shopping, logistics is in urgent need of innovation. Thelag of logistic level makes that the logistics distribution has become the key factor restricting thedevelopment of e-commerce mode. In order to improve the efficiency of logistics distribution, thestudy on vehicle routing problem (VRP) of logistics distribution occupies an important position,which realizes intelligent management of distribution operations.Due to the fresh agricultural product distribution with perishable characteristics, make use ofthe Vehicle Routing Problem with Time Windows (VRPTW) to establish its mathematical model.The VRPTW is NP-hard, and the efficient solution algorithm exists unlikely. This thesis hasconducted the research to the fresh agricultural product distribution path optimization problems,mainly including the following aspects:(1) A review of research related on the vehicle routing problems. Summarize the domestic andforeign evaluation on classification of VRP and its path optimization criteria, and comparativelyanalyze the advantages and disadvantages of various algorithms and their application of occasions.(2) In view of the basic ant colony algorithm’s defects of easily falling into local optimalsolution and slow global convergence speed, propose an improved ant colony algorithm to solve theVRPTW. The algorithm is applied to VRPTW benchmark problems; the experimental results showthat the algorithm is effective.(3)By the advantages of fast convergence speed of ant colony algorithm and local optimizationability of large neighborhood search algorithm, propose a hybrid ant colony algorithm to solve theVRPTW. The algorithm uses the ant colony algorithm parallelism and positive feedback to generateinitial solution, and then uses the large neighborhood search algorithm to exploit the neighborhoodstructure of initial solution by across adjusting the initial solution in the local optimization process.The VRPTW benchmark problems demonstrate the effectiveness of the algorithm.