A post-classifying fuzzy-neural and data-fusion rule for job scheduling in a wafer fab - a simulation study

Abstract

This paper proposed a post-classifying fuzzy-neural and data-fusion rule to improve the performance of job scheduling in a wafer fabrication factory (wafer fab). The proposed rule is a hybrid (fusion) of two well-known fluctuation smoothing rules – FSMCT and FSVCT. Several ways of data fusion [including normalised sum (NS), normalised product (NP), condensed normalised product (CNP), weighted normalised product (WNP), and dynamic weighted normalised product (DWNP)] were applied for this purpose. Besides, in order to enhance the scheduling performance of the rule, the post-classifying fuzzy back propagation network (FBPN) approach was applied to improve the forecasting accuracy of the remaining cycle time. To evaluate the effectiveness of the proposed methodology, a production simulation was carried out. According to the experimental results, the proposed methodology outperformed some existing approaches by simultaneously reducing the average cycle time and cycle time variation.