| |
| Semiconductor wafer fabrication is a very complicated manufacturing system with characteristics such as reentry of lots into machines, alternative machine with unequal capacity and shifting bottlenecks. The optimization and scheduling of semiconductor wafer fabrication has long been a hot research field. Bottleneck is the key factor to the wafer fabrication which has essential influence on the throughput rate, cycle time, time-delivery rate, etc. The scheduling policies on bottleneck machines have significant impact on cost decreasing, achieving throughput target and improving production-resource scheduling and therefore need to be optimized. In this paper, considering the characteristic information concerning bottleneck and whole production line, a rolling horizon bottleneck prediction method is proposed through predict the wait time of the lot in machines and the machine load. Then, based on the above prediction bottleneck, we proposed a novel rolling horizon dynamic bottleneck-based scheduling algorithm. The algorithm analyzes the key factors which influence the bottleneck shifting and the upstream and downstream scheduling policies of bottleneck. Applied to a simulation wafer fabrication, our proposed bottleneck prediction method and bottleneck-based scheduling algorithm improve the cycle time, machine utility apparently compared to other heuristic scheduling method |
|
| Keywords:control theory and control engineering; semiconductor wafer fabrication; dynamic bottleneck prediction; bottleneck-based scheduling algorithm |
|
