A nonlinear scheduling rule incorporating fuzzy-neural remaining cycle time estimator for scheduling a semiconductor manufacturing factory—a simulation study

Abstract

A nonlinear scheduling rule incorporating a fuzzy-neural remaining cycle time estimator is proposed in this study to improve scheduling performance in a semiconductor manufacturing factory. The proposed scheduling rule is modified from the well-known fluctuation smoothing rule with three treatments. At first, the look-ahead self-organization map-fuzzy back-propagation network approach in our previous study is used to estimate the remaining cycle time of every job in the semiconductor manufacturing factory. Subsequently, the release time and remaining cycle time of a job are both normalized to balance their importance in the fluctuation smoothing rule. Finally, the normalized release time is divided by the normalized remaining cycle time to obtain the slack. In this way, the proposed scheduling rule becomes a nonlinear one. To evaluate the effectiveness of the proposed methodology, production simulation is used to generate some test data. According to experimental results, the proposed methodology outperformed many existing approaches in reducing both the average cycle times and cycle time standard deviations. The advantage was up to 41% over the basis p-FS policy when the cycle time standard deviations were to be minimized.