Research On Dynamic Batch Scheduling For Semiconductor Furnace Area Based On Petri Nets

Furnace area is one of the four significant zones in semiconductor manufacturing system. Long cycle time, batch processing mode, and a great mount of WIP make it one of the main bottlenecks in system. This paper researches on the dynamic scheduling in furnace area.Semiconductor manufacturing systems wafer manufacturing system is characterized by great scope, diverse equipment, re-entrant manufacturing process and volume-produce multi-style. A hierarchical modeling method based on object-oriented hybrid Petri nets (OHPNs) is proposed. It divides the modeling process of furnace area into device, manufacturing unit and product area. A Petri net model for furnace area is constructed.The dynamic scheduling system for furnace area is made up of OHPNs model, monitor, scheduling algorithm and database. When monitoring delay, breakdown or preventive maintenance etc. and out of the default standard, the system would reschedule with dynamic scheduling algorithms according to current state.A scheduling algorithm named weighted cost rate heuristic for single furnace multi-product and a two-level scheduling algorithm based on PSO and WCRH algorithm for non-identical batching device group are proposed. They make use of the future arrivals information. Results show that these algorithms improve the weighted earliness/tardiness performance. Based on dynamic Petri net scheduling system, the results of simulation show that it is effective in improving the whole furnace area's performance in on time delivery.