Attribute selection algorithm of data-based scheduling strategy for semiconductor manufacturing

Abstract

In today's digital and information-based manufacturing environment, data are basic elements for almost every production control and management activity. This paper focuses on production data processing based on attribute analysis. There are thousands of attributes in semiconductor manufacturing. However, some of them are irrelevant and/or redundant to some optimal production control and management issues. It is hard to decide which attributes should be considered as input references. Rational attribute selection may lead to an accurate scheduling strategy and finally exerts a positive impact on the performance of the whole production line. This is the motivation of this work. Its goal is to investigate which attributes play the key roles in the manufacturing scheduling according to a specific performance criterion. A genetic algorithm-based selection approach for feature production attributes is proposed. Its prediction accuracy is verified via a practical wafer production line.