Supply Chain Networks Design under Demand Uncertainty: A Fuzzy Multi-Objective Optimization Approach

• Main Authors: Yuan, Tzu-Wei, 袁子偉
• Other Authors: Chen, Cheng-Liang
• Format: Others
• Language: zh-TW
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/22913833179330017912

碩士 === 國立臺灣大學 === 化學工程學研究所 === 91 === This research aims at studying the design strategy for selecting the total numbers and their locations of warehouses and distribution centers in a supply chain network. We consider modeling a multi-product, multi-echelon, and multi-period supply chain network. The network comprises a number of manufacturing sites and a number of customer zones at fixed locations, and a number of candidate warehouses and distribution centers of unknown locations. The transportation modes are economy-of-scale modes. And we use several different scenarios to deal with customer demand uncertainty. The decisions to be determined include the numbers, locations, and capacities of warehouses and distribution centers to be set up, the transportation links that need to be established in the network, and the quantities of transportation and production for the products. In traditional supply chain management, minimizing total cost as a single objective is often the focus when considering integration of the whole supply chain network. In this research, we simultaneously consider four design objectives: the total costs, robustness to uncertain demand, local incentives, and total transport time. The problem is then formulated as a Multi-Objective Mixed-Integer Non-Linear Programming mathematic model (MO-MINLP). The fuzzy multi-objective optimization method is adopted to solve the problem. Two popular t-norms, the product and the minimum, are respectively used for implementing the fuzzy intersection during optimization. It is found that using the product operator may cause an unbalanced solution between all objectives, while the minimum operator may result in multiple solutions. We thus propose a two-phase method which can combine advantages of these two t-norms. The minimum operator is used in phase 1 to find the least degree of satisfaction, and the product operator is then applied in phase 2 with guaranteed least membership value for all fuzzy objectives as additional constraints. Finally, one numerical example is supplied, demonstrating that the proposed two-phase method can provide a better compensatory solution for multi-objective optimization problems in a supply chain network.