Distribution Problem And Its Simulation Based On Road-Net Model With A GIS Rich Network Properties

In recent years, modern logistics, known as "the third profit source", has a huge effecton the transportation cost of the enterprise and the satisfaction of the customer. So, modern logistics catches the attention of the researchers and the enterprise. Distribution problem is the last but the most important part of the modern logistics, occupying a large part of the enterprise logistics expenses and the direct contact with customers. So, optimizing the logistics distribution to decrease the cost of the whole logistics system and to improve the level of logistics services plays a key role in the whole process. Introducing GIS technology can effectively analyze and dealwith the fusion problem in route optimization of the distribution, such as, the approach to each path restrictions, the number of lanes, the congestion and multimodal distribution of the vehicle.The combination of GIS technology and logistics distribution contributes to the logistics decision, the improvement the efficiency of distribution, the control of the distribution costs, enhancement of customer’s satisfaction, and the long-term scientific development of the enterprise.To address the deficiencies in traditional logistics problems, this paper proposes a framework of GIS rich network distribution network model, using the N-order adaptive partitioning algorithm for neighbor distribution outlets set division classification. Firstly, the GIS rich-net integrates in logistics distribution model, and build the framework of distribution model with multiple distribution centers and more distribution outlets, and adress abstract problem in road network and reduce distribution model adaptation restrictions. Secondly, we add time window, vehicle type, vehicle wear and driver costs to the distribution model. Then, the large-scale distribution network onthe established Distribution lines adapts N order nearest adaptive algorithm, to classify the distribution network. By building distribution algorithm, logistics vehicle route gets its optimization. Finally, the simulation experiment results show that the model of established distribution lines and distribution algorithm design is good. Firstly, we summarize the development of logistics and distribution of the VRP model achieved and some achievements. The solving of the VRP problem with exact algorithm and heuristic algorithm arepointed out.GIS-rich network attributes model that combines the shortest path problem multiple weights for the actual road network when abstraction for managing spatial data and dealing with unique features. Also for solving the shortest path problem corresponding algorithm described in detail, we compared their characteristics of different optimization algorithms to determine the applicable conditions of various algorithms. VRP problem for large-scale distribution outlets in the collection, the classification algorithm is proposed and described in detail in several currently used classification algorithm and application conditions.Secondly, the multi-depots logistics distribution problem is a particular combination of transportation and vehicle routing problems, in which has to be loaded into different customers within vehicles, and then delivered with minimum cost. In this paper, A GIS rich network model is built firstly and the weights between the roads are acquired by GIS software; next, the idea of the evolutionary tree classification is used to solve the nonlinear cost variables in the distribution model. Extensive computational results show that the proposed algorithm is able to solve the optimality for a large number of small-size instances. The distribution model is introduced into the rich network attribution, but it considers little factors.These factors play an important role in the following chapters establish fusion model too.Thirdly, Vehicle Routing Problem with Time Windows (VRPTW) is a well known NP-hard problem. A typical way to solve this problem is to devide the customers into groups first, and then construct optimal routes for each group or groups. Conventionally, customers are partitioned according to the spatial distance by GIS, while ignoring their time window requirements. In the paper, spatial and temporal characteristics of customers are considered and a spatiotemporal distance based criteria for customers partitioning are proposed. To improve the initial solution, N order shortest neighbor and GA(Genetic Algorithm) are used. Extensive computational results show that the proposed algorithm are able to solve the optimality for a large number of small-size instances, and consistently outperforms previous algorithms from the literature on large-size and real-world instances.Finally, we makes a brief summary of the whole dissertation, presents the achievements and shortcomings as well. What’s more, researcher propose some advices for further research.