Abstract
In the semiconductor industry, mean cycle time is usually one of the key perfor-mance measures used by fabrication facilities to maintain their competitive advantage. As the machine in the photolithography area is more expensive than that in other areas, photolithography is usually a bottleneck area that significantly affects mean cycle time performance. Because of multiple production processes and their respective re-quirements and routings (e.g., precision requirements in alignment), multiple work-stations are present in photolithography areas. So, how to balance the workload among machines and how to reduce the setup time for changing the temperature groups are the important factors to reduce the mean cycle time. In this study, a two-stage dispatching rule for photolithography areas is proposed. In the first stage, a heuristic method for balancing workloads is developed, in which the machine with the lowest current and waiting workload is used. In the second stage, the heuristic rule is derived based on the concept of the smallest setup time for a sequence dependent setup time problem. Numerical studies show that with different scenarios, the two-stage dispatching rule can reduce the mean cycle time from 7% to 10% compared with cur-rent practice. Finally, according to the simulation result in different scenarios, we would discuss about the relationship among mean cycle time, arrival rate, and differ-ent scenarios.