Ant Colony Optimization Based Scheduling for a Semiconductor Wafer Fabrication Facility with Bottleneck Stations
Li,Fei Qiao,Xiaoyu Tian,Qidi Wu
DOI: https://doi.org/10.1109/ical.2009.5262868
2009-01-01
Abstract:In this paper, it is firstly noted that the scheduling of a semiconductor wafer fabrication facility (fab) has its special complexities, such as large scale, a mixed of different processing types, unbalanced production facilities, short-term efficient scheduling period, non-zero state at the decision point time, etc. Secondly, the scheduling model for a wafer fab, composed of a machine set, a task set, a processing time set and the optimized objective function focused on the movements and delivery performance of the jobs, is built with fully consideration on its characteristics. Thirdly, an ant colony optimization based scheduling algorithm (abbreviated as LSA-WF) is proposed. The main idea of LSA-WF is the scheduling of the bottleneck machines of a wafer fab is considered first and foremost, the scheduling of the batch processing machines with recipe constraint is in the next place, the scheduling of machines by wafer or by lot is the third, and the scheduling for the batch processing machines without recipe constraint is the last. The same type of machines is ranked by their workloads in the descending order. Then the simulations on a real wafer fab are used to verify and validate the proposed method. The simulation results show that the proposed method is superior to the common rules (such as FIFO, EDD, SRPT and CR) with more movements of the jobs and higher machine utilities. Finally, the summarization of the research results is given.