Application of Petri Net and Genetic Algorithm in Optimal Scheduling of TFT-LCD Manufacturing Process

• Main Authors: Hsi-Tsun Pao, 包溪圳
• Other Authors: Chun-Liang Lin
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/74643823797992026723

碩士 === 國立中興大學 === 電機工程學系所 === 98 === How to develop the optimal strategy of job scheduling control for factory automation (FA) is the major concern of computer integrated manufacturing in TFT-LCD (Thin Film Transistor Liquid Crystal Display) industry, which complicated production processes and long lead times are the major features. In general, the production processes in TFT-LCD industry can be classified as array, cell, and module. Each production process should perform within its dedicated workshop and thus form a 3-layer supply chain that is capable of fulfilling all customer orders. Production module simulation with Petri net, which required automatic definitions of production lines, has allowed TFT-LCD production module approach to real production processes. And, it is possible to obtain better or the best scheduling rule by applying genetic algorithms (GAs) to improve production efficiency. In this study, we have constructed a G-P scheduling system (G-PSS), which equipped both of GA and PN, developed in the Matlab environment. The design scheme simulates and evaluates all possible production scheduling plans, then informs the workshop with the best solution. After properly setting two major variables: working time and rate of machine usage, applications of this system are expected to increase the output of production lines, the capacity to deal with high variety products and utilization of machines. G-PSS is, therefore, able to reach the goal of cost reduction by optimizing the production schedule and deposition of production resources.