A GENETIC ALGORITHM FOR SOLVING STOCHASTIC PROGRAMMIN MODEL OF FACILITY LOCATION PROBLEM IN 3-ECHELON SUPPLY CHAI OF TFT-LCD INDUSTRY

• Main Authors: Sin-Yi Liu, 劉欣怡
• Other Authors: Kung-Jeng Wang
• Format: Others
• Language: en_US
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/69892486138180334636

碩士 === 國立東華大學 === 企業管理學系 === 94 === Many Taiwan TFT-LCD panel manufacturing firms are undergoing new plant deployment plans; however, they suffer from investment and market risks at the same time. Therefore, the decisions about building new generation TFT-LCD plants become an important issue to success. Furthermore, a typical TFT-LCD supply chain (SC) network is a three-echelon structure, so a manager must determine how to decide the best distribution strategy to application domain customers in an effective way. In this study, we investigate the two-echelon and three-echelon SCs by scenario-based stochastic programming approaches in which the issues of selecting facility location and deciding delivery problem, accounting for uncertainty in demand. This is accomplished by integrating genetic algorithm (GA) and greedy algorithm to solve the stochastic programming model. Experiments reveal that the proposed GA-Greedy algorithm can solve successfully the problem in a timely way. The Experiment outcomes were also compared to the optima of the mathematical programming model. Sensitive analysis confirms the robustness of the proposed algorithm under a variety of the algorithmic parameters.