Research On Three-hierarchy Supply Chain Coordination Under Fuzzy Demand Based On Bi-level Programming

The supply chain coordination has become the core strategy of the supply chain management. From the literature reading, the research of the supply chain coordination mainly concentrated on two-hierarchy supply chain coordination model and the three-hierarchy supply chain coordination model mainly concentrated on ’1:1:1" style, that is, one supplier, one manufacturer and one retailer. And the research of more meaningful "M:1:N" style is still not enough."M:1:N" style supply chain means the system is consisted of M suppliers, one manufacturer and N retailers. And the current research mainly concentrated on the common product the supply chain provided, and the description of innovative products which has the features of short life cycle and soon technical updation is not engough either. Based on this situation, the paper made a research on The supply chain system which provided innovative products and was faced with fuzzy marker demand, and the paper proposed a "M:1:N" style three-level supply chain coordination decision problem:Firstly, the paper reviewed the corresponding domestic and foreign research from three aspects, that is, supply chain coordination mechanism, supply chain structure and the market demand environment the supply chain lying to, and found the insufficiency of the current research and proposed a new coordination mechanism:the combination of price discount with coordination development contract.Secondly,according to the dynamic game of the members and the fuzzy market demand, the paper setted up the three-hierarchy supply chain no-coordination mechanism model and the model under the fuzzy demand environment by the tactics "coordination development" and price discount contract with the thought of bi-level programming and fuzzy sets theory. The whole model is divided into two level layers. The upper level planning’s objective function is the expected profit of the core enterprise, that is manufacturer, and the lower level planning’s objective function is the expected profits of the non-core enterprises, that is the suppliers and the retailers. The relationship between the upper and the lower plans is:the upper level layer planning’s objective function and constraints not only depend on the upper level layer decision-maker’s control variables but also depend on the lower level layer decision-maker’s optimal solution; and the optimal solution of the lower level layer planning is also influenced on the upper level layer’s variables, which helps to form an interconnected supply chain decision-making system. The Model solution can obtain the expected profits of these two decision modes and the values of wholesale, sales and purchase prices, the price discount rate, order quantity, subsidies for development and investment for development.Then, according to the specjalties of model multidimellsional parameters, programe composed explosion and hard solving, the paper applied Chaos Search algorithm for the model solving. Chaos Search algorithm has the interval ergodic feature, and without knowing the specific function of the lower level’s feedback, just to know each variables’internal value, will facilitate to solve the bi-level programming model.Finally, the paper setted up a tested example to verify the feasibility of the model. And the result of the example showed that the manufacturers can coordinate the interest among the members by the application of the way of "coordination development " and "price discount contract" and realize the Pareto improvement.