The Application of the Improved RBF in the Semiconductor Manufacturing System Feeding Control

摘要

According to the complexity of the semiconductor manufacturing processes, a RBF neural network improved by GA is put forward to apply in the semiconductor production modeling and predict the line performance indexes. The improved RBF algorithm can play better role in dealing with the dynamic real-time data from the line, building predictive models to describe line dynamic behavior, and presenting future outputs based on the current inputs accurately. In this paper, the improved RBF algorithm and the traditional RBF neural network are performed respectively to realize the line modeling simulation. Finally, the simulation result shows the effectiveness of this improved algorithm.