Research On Optimization Method For Vehicle Routing In Logistics Distribution

In recent years, China's logistics industry has achieved great development, but as compared with foreign developed countries, there are still large gaps. In all the links of logistics, logistics distribution plays a key role in increasing profits of logistics enterprises. As part of solutions to the technical problem of logistics distribution, vehicle routing problem(VRP) is getting more and more attention in academics and enterprises.VRP is a typical NP-hard problem. Even in the relatively small size of the customers, getting the solution also is difficult. Therefore, studying under all conditions for the effective VRP algorithm appears to be particularly important. Judging from the current situation of the study, although the study on VRP has got much attention, but still not on the actual situation facing the VRP-depth study. The VRP research results are rather scattered, the depth and breadth of which are unable to meet today's rapid development of the logistics industry.The textual main work and innovation points include:Firstly, a branch and cutting algorithm is designed to solve CVRP. The algorithm uses multiple cutting surfaces, the improved branch rules, the nodes selection and deletion rules, the cutting surface updating rules, and so on. Through the comparison of domestic and international examples, the algorithm proposed in the dissertation embodies the performance superiority.Secondly, the performances of many intelligent algorithms to CVRP are analyzed and compared. These algorithms include genetic algorithm, annealing algorithm, tabu search algorithm. In the dissertation, the advantages and disadvantages of the various algorithms in dealing with the specific issue of the VRP are analyzed, and the most suitable frame of intelligent algorithm is identified for the optimization of VRP.Thirdly, the concept of core path is proposed in the dissertation. Based on core path, a tabu search algorithm is designed to solve CVRP. The standard database is used to test the ability of the algorithm optimization, computational speed and stability. Compared with the similar algorithms, the algorithm presented in the dissertation has more advantages.Fourthly, various non-full load VRPs are systematicly studied in the dissertation. These VRP conditions include: single-depot VRP with pick-ups or deliveries; single-depot VRP with with pickups and deliveries; single-depot open VRP; multi-depot VRP with pick-ups or deliveries; multi-depot VRP with pick-ups and deliveries.Fifthly, the real-time VRP is studied, which can be applied to the situation when vehicles have started from the distribution center and the customer's requirements suddenly changed. The emergency scheduling problem is the preliminary discussion to the more complex VRP, and it is closer to the actual application.Sixthly, the logistics distribution equilibrium principle is proposed. The bilevel programming model of VRP problem with logistics distribution center location is presented in the dissertation. Such work is the expansion of VRP in integrated logistics environment.Seventhly, CVRP and the automated warehouse picking problem are compared. The research results are applied to the actual enterprise and the application achieves good results.