As an emerging service industry of prospects and tremendous potential market, modern logistics has become the source of power to ensure the sustainable development of the national economy in a high starting point, and has expanded into integrated logistics system with scientific technology and modern management as the pillar.By optimizing the logistics system, realizing the resource optimal allocation and element optimal combination, it will form divergent effect, promote and stimulate new economic growth, achieve the target of resource conservation, environment-friendly and logistics optimization.The optimization of distribution center location and transport routes are two important sub problems in logistics system. The previous studies are mostly independent analysis of the two problems, resulting in being difficult to achieve the overall optimization effect. At the same time there are some blind spots. Therefore it is necessary to combine the logistics characteristics to conduct a study focusing on the combination of Location-Routing Problem. In this dissertation, the home and abroad research status were reviewed and analysed, and through comprehensive consideration and analysis, some of the important issues are studied on which to establish the model and design algorithm for application.As the key link platform, distribution center location is very important for logistics system optimization. Based on the definition of the connotation of the distribution center, the discrete location methods are introduced according to the need of research, improvement minimum envelope clustering model are designed to solve location problem. The node allocation principle and algorithm procedure is optimized for the model based on gradual covering and the minimum envelope clustering model on the basis of sound judgment considering the conditions of each node’s service time and spatial distribution.The problem of Location-Routing research system under static environment has been relatively matured, but some fields still exist a research gap. Based on forward logistics distribution research and fully consideration of the characteristics of reverse logistics, this dissertation builds model to sovle the optimization problem and selects supermarket distribution example for test.With further research, scholars have started to pay attention to Location-Routing problem in dynamic environment. This dissertation has defined the problem more realistic, regarding those uncertain factors like route network performance as problems. This kind of problem is studied based on the dynamic change in route network performance without considering that service facilities may change location. With the idea of real time optimization, the optimization problem in dynamic environment is converted into determined static problem by stages.Based on the theoretical study, this dissertation selects two timeliness strong examples of manufacture distribution and cold chain distribution to test the dynamic environment problem, and chooses the minimum envelope clustering model combined with time windows requirement to locate, and optimizes vehicle routing on the basis of maximum customer satisfaction through the analysis of customer satisfaction model, so as to achieve the profit of enterprises and customers.Considering the complexity of the problem, this dissertation applies the method, decomposition first and integration later, to design a solution so as to ensure its quality. The problem is divided into two sub-problems, thinking over solutions for the sub-problem of location and the sub-problem of route optimization firstly. The sub problem is cohesive rather than independent relationship, and then it is going to think about the integrity of problems so as to gain the final solutions.In this dissertation, tabu search algorithm is used to solve the problem with reverse logistics under static environment. In addition, the dissertation solves one location-routing problem with time windows and another problem of heterotypic vehicles in dynamic environment through using hybrid genetic algorithm, it conducts the example test to obtain a more satisfactory optimization results in a shorter time, which shows the advantages of the algorithm. |