A Study of Dispatching for Parallel Machines in TFT-LCD Assembly Processes

• Main Authors: Chi-TingFu, 傅祺婷
• Other Authors: Yeu-Shiang Huang
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/40528401324445808575

碩士 === 國立成功大學 === 工業與資訊管理學系專班 === 101 === Thin Film Transistor Liquid Crystal Display (TFT-LCD) has gradually replaced the traditional cathode ray tube (CRT) display in recent years because of its compact size and power-down feature, which leads to several challenges that manufactures have to face, such as a small amount of production volume, product diversification, shorter product life cycles, and decreasing order delivery times. This study deals with the dispatching problem for parallel machines in TFT-LCD assembly processes. Since orders from clients and material from suppliers would constantly change with time, which results in the requirement of flexibility for scheduling of orders, the TFT-LCD manufacturer has to develop a more appropriate production planning and scheduling system to solve the production problem. Because the non-equivalent parallel machine scheduling problem is NP-hard, and it has high complexity and is not easy to be solved and moreover, when the problem scales bigger, it becomes un-efficient to use traditional optimization methods, this study is thus based on the genetic algorithm to solve the unrelated parallel machine scheduling problem through the setting of a variety of performance measurement indicators. In the end, we compare the solutions from the Genetic algorithm with these from LINGO and analyze them, and the results show that the solutions from the Genetic algorithm are not worse than these from LINGO but with shorter computation time. Therefore, it saves lots of time and has an acceptable quality to have a solution through the Genetic algorithm.