Optimized fuzzy-neuro system for scheduling wafer fabrication

Abstract

Optimized fuzzy-neuro system is proposed to improve performance of scheduling jobs in a wafer fabrication factory(WFF). Fluctuation smoothing rules are normalized. Remaining cycle time of a job is accurately estimated by applying self-organization map with fuzzy back propagation network (SOM-FBPN). Division operator is applied instead of traditional subtraction operator to enhance responsiveness of fuzzy rule. Production simulation using proposed methodology outperformed existing approaches in reducing average cycle time and cycle time standard deviation.