Finite capacity planning algorithm for semiconductor industry considering lots priority

摘要

A finite capacity planning heuristic is developed for semiconductor manufacturing with high-mix low-volume production, complex processes, variable cycle times and reentrant flows characteristics. The proposed algorithm projects production lots trajectories (start and end dates) for the remaining process steps, estimates the expected load for all machines and balances the workload against bottleneck tools capacities. It takes into account lots' priorities, cycle time variability and equipment saturation. This algorithm helps plant management to define feasible target production plans. It is programmed in java, and tested on real data instances from STMicroelectronics Crolles300 production plant which allowed its assessment on the effectiveness and efficiency. The evaluation demonstrates that the proposed heuristic outperforms current practices for capacity planning and opens new perspectives for the production line management.