Abstract
Abstract As high-tech industry including semiconductor manufacturing industry breaks into a tiny-profit age, many studies examine lot size variation to decrease cycle time and cost. High-tech industry including semiconductor and TFT-LCD manufacturing industry has similar manufacturing properties, this study aims to examine and monitor the lot size variation to study its impacts on performance indexes. Based on a fab simulation model, in this research also use data mining approach to find out critical machines in the manufacturing process and extract potential useful patterns. Thus, we will try to suggest better WIP level arrangement on the basis of the results. We have employed this approach in a real TFT-LCD fab as an empirical study for validation. Key words: Semiconductor manufacturing, TFT-LCD manufacturing, Data mining approach, WIP level