Optimization Models And Algorithms Of Distribution Network Under Fuzzy Environment

Distribution network optimization problem is a key topicin the supply chain management, and the performance of which affects the enterprise’s benefit and customers’interests. The topic is studied extensively in the operational research, combinatorial optimization and management science research.In this paper, the vehicle routing problem (VRP), location-routing problem (LRP) and location-routing problem with simultaneous pickup and delivery (LRPSPD) were studied as case study. Since some unknow information often appear in real distribution system, the problems with uncertain may not be solved by research theory and method from determine problems, so the research on uncertain problem has great significance.We used fuzzy logic to deal with the uncertain problem, and the VRP, LRP, LRPSPD under fuzzy environment were studied.For VRP under fuzzy uncertain environment, we selected the VRP with fuzzy demand (VRPFD) as case problem, and in which the capacity of vehicles was considered. The hybrid scatter search algorithm (HSSA) was developed, and a global real-time adjustment strategy was put forward. Finally, the validity of the developed algorithm and the proposed strategy were verified. For LRP under fuzzy uncertain environment, we selected LRP with fuzzy demand and fuzzy travel Time (LRPFDFTT) as caseproblem, and in which time window of customers and the capacity of vehicles were considered. Couple the combination effects of fuzzy demand and fuzzy travel time, a chance-constrained model with change-reward was proposed and the hybrid simulated annealing algorithm (HSAA) was used. Finally, the effectiveness of the proposed model and the used algorithm were verified. For LRPSPD under fuzzy uncertain environment, we selected the location-routing problem with simultaneous fuzzy pickup and fuzzy delivery (LRPSFPFD) as case problem, and in which the capacity of vehicles was considered. A three dimensional-index MIP model was set up according to the problem features, and the correctness of the model was verified through the software LINGO solving simple numerical example. A variable neighborhood scatter search (VNSS) was proposed and the effectiveness of the algorithm was verified. For the problem application, we selected the B2C fuzzy distribution networkoptunization problem as case problem, a fuzzy bi-leveloptimization model coordinating fuzzy-static facility location and allocation problem (LAP) and certain-dynamic vehicle routing problem(VRP) was developed. A genetic algorithm embedded random algorithm and tabu search algorithm was designed. Finally, the effectiveness of the developedmodel and the designed algorithm were verified.