The Research Of A Robust Optimization Model For APP In A Supply Chain Under Uncertainty Of Demand And Cost

The life of product is shorter, the competition is more fierce and the demand is more uncertain in today’s market. With the network structure of supply chain, these factors bring about uncertainties in the supply chain. It means risk to supply chain, and maybe destroy the chain. So we need to use the optimization method to reduce the risk of uncertainty, and minimize the uncertainty, in this way, we can improve the competitiveness of the enterprises. In production management, APP is an important technical planning; In the field of planning, it is between the long-term planning of broad decisions and the detail decision in short-term planning. Therefore, the research of the APP in supply chain is of great significance.In this paper, we study such APP:with a supplier, multi-manufacturers and multi-client area. Raw materials required by the manufacturer are supplied by the supplier, and every manufacturer can provide every client area with products. Each manufacturer can adjust its labor force by hiring or firing, and change productivity in the form of outsourcing or overtime. We first to define the concept of APP, then we classify the uncertainties by different standards. In this article, we select two kinds of uncertainty-the demand and cost, and use scenario analysis to describe the uncertainties, and build APP model with the robust optimization method. Respectively, the functions to minimize the loss and maximize customer satisfaction are constructed. We consider two objections simultaneously, the first one aims to minimize losses in the chain, and the second one aims to maximize the products provided to customer. This paper gives the way to handle the problem of multiple targets, namely the technique of normalization. The multi-objective model can be changed into single objective model. In order to validate the feasibility and effectiveness of the model, this article designed the case, and solved it based on LINGO software. In the end of the article, we analyzed the robust of model based on the results of LINGO software solution, not only analyzes the change trend of the objective function with the change of risk parameter, also analyzed the change trend with the change of weight parameter. According to the results of this paper, we know the model is robust. In order to validate the importance of penalty function, this paper also gives a graphic solution without penalty function in the model, and compared with the graphic with penalty function in the model. The result shows that the robust optimization model with penalty function is better. At the end of the paper, we made the trade-off graph between objective function1and the objective function2, policymakers can decide the trade-off between the loss and fill rate of demand according to the actual situation of enterprises.This paper constructs the APP robust optimization model, by solving a case, not only proved feasibility and effectiveness of the model, but also gave the trade-off curve between the objective functions, this has some reference values to enterprises.