Multi-depot Open Vehicle Routing Problem Based On The Integration Of A Variety Of Factors

Vehicle routing problem is important research areas in the field of the logistics supply chain optimization and combinatorial optimization. Its research priorities is transforming from single factor to multi-factors and vertical integration, or basing on specific industry and business model.In this paper, Based on the dairy industry and Combined with the current characteristics of logistics and chemical products. We research the complex logistics problems considering the following factors:multi-center, multi-product, heterogeneous vehicles, exclusive products, products’transport conditions, vehicle sharing and leasing, cargo transfer among the warehouses, fuzzy customer demand. In the research methods, the gradual deepening method is proposed. That is to say, we split the problem into three sub-problems and gradually add factor based on the pre-study. In the ideology, the "segmentation processing and overall optimization" is adopted to solve the sub-problemsFirstly, to solve the insufficiency of transportation capacity in a company and the imbalance of transportation capacity among its decentralized distribution centers in each decision-making process, a mixed-integer programming model was developed for multi-depot open vehicle routing problem (VRP) with time windows based on vehicle leasing and sharing. By introducing a virtual distribution center, the multi-depot VRP was transformed into a single depot VRP. Then a hybrid genetic algorithm was proposed by combing scanning algorithms with C-W saving algorithm, as well as optimizing vehicle routing and vehicle scheduling together. Finally, the validity of the model and algorithm was demonstrated by the real data from logistics branch of Chongqing Tianyou Dairy Co., Ltd... The results show that the model and algorithm performs better than company’s existing solution in the following aspects:total travel mileage, total cost, and time en route of delivery vehicles. In addition, the sensitivity analysis of model parameters and the convergence analysis of the algorithm were done, and the results show that the algorithm has better performance.Secondly, considering the product’s diversity, mutual exclusion, and heterogeneous vehicles that can delivery some products because of transport properties. A nonlinear programming model was developed for the multi-depot open vehicle routing problem with heterogeneous vehicles and exclusive products. Then, we proposed a combination of genetic and two-phase heuristic algorithm, according to the principle of "fragmentation process, the overall optimization". Finally, the validity of the model and algorithm was demonstrated by the real data from logistics branch of Chongqing Tianyou Dairy Co., Ltd.. The results show that the model and algorithm performs better than other existing solution and prove that a mixed multi-product distribution among the various products is much lower than the cost of transportation alone. In addition, the sensitivity analysis of model parameters and the convergence analysis of the algorithm were done, and the results show that the algorithm has better performance.Finally, adding the factors:cargo transfer among the warehouses and fuzzy customer demand. I study the integration of cargo transfer and vehicle sharing among the warehouses. Meanwhile, a nonlinear programming model was proposed to explain the problem with mathematical language.