关键词: 供应链网络, 成本优化, 收益优化, 双目标混合整数规划, 增广ε-约束方法

Abstract: Compared with the developed countries, the cold chain logistics started relatively late in China. Nowadays, there still exists some problems in cold chain logistics in China. For instance, facilities are not complete, cold chain management is not perfect, cost is high, profits are low, and so on. Especially, the supply chain network of dairy products with short-life-cycle is a special class of cold chain logistics. Since the dairy products with short-life-cycle are perishable, we need not only to make sure of the quality of dairy products, but also to send them to the right place in the fastest way, in the shortest possible time. Meanwhile, the dairy products with short-life-cycle have some characteristics. For example, the customer demand is uncertain and difficult to predict, product variety, the small amount of product's order, etc. Thus, the requirements for dairy products supply chain network management are higher, and subsequently, the corresponding cost is relatively high.One of the goals of researches on supply chain network is to minimize the total cost and maximize the total revenue at the same time under the condition of the whole supply chain network satisfying the customers certain demand. In this paper, for short-life-cycle dairy products which have many varieties and are difficult to store as well as have some special requirements in the process of transportation, the related cost and revenue optimization problems for the four levels of dairy products supply chain network with "supplier-manufacture-distributioncenter-retailer" are studied. Under the condition of satisfying the customers demand for some products, a double-objective mixed-integer programming model is established, which can ensure that the cost and the revenue of the dairy products supply chain network could obtain the optimal value at the same time. Moreover, in order to describe the product's practical inventory for retailers effectively, a new nonlinear function is introduced in this paper to define the dairy product's inventory. By making use of the structure characteristics of the proposed model, a new augmented ε-constraint method is designed to solve the model. The empirical analysis results show that the proposed model could not only provide optimal decision-making scheme such that the cost and the revenue of the whole supply chain network could achieve the optimal value at the same time, also give the related optimal decision schemes for the products production, transportation and so on. Furthermore, the designed method is more promising than the standard ε-constraint method. The related research results have important potential guiding significance for studying how to optimize the cost and the revenue of supply chain and improve the customer's satisfaction, solving the general double-objective optimization problems and so on.

Key words: supply chain network, cost optimization, revenue optimization, double-objective mixed-integer programming, augmented ε-constraint method