Research On Supply Chain Network Design Under Uncertain Demand

Under the conditions of modern market economy, with the increasingly fierce global competition, the continuous complex supply chain network, the shortness of product lifecycle and the growing diversity of customer needs, manufacturers gradually formed a multi-vendor, multi-plant, multi-distribution center, multi-customer distributed management mode of supply chain. Faced with increasing competitive pressures, enterprise has to consider a variety of uncertainties and to re-design their supply chain networks in the case of limited resources, in order to optimize their processes and improve the overall performance.In this paper, the structure of supply chain network is analyzed first. Then, we point out the influencing factors in the supply chain network design and the main principles to be followed. The fuzzy optimization design of supply chain network under single-source, multi-product, multi-stage environment is studied. According to Zadeh's possibility theory, the model can be converted to deterministic model by converting chance constrains to their respective crisp equivalents.It's is difficult to solve the model by traditional method. Hence, a hybrid genetic algorithm which combines priority encoding and integer encoding is used to solve the problem. The encoding and decoding method, the initialization of population, the fitness function, the design of genetic operators are discussed, At last, the feasibility of GA to solve such problem is verified by a numerical example.