关键词: 半导体生产线, 并行批处理机, 优化, 调度, 嵌套分区, 总加权拖期交货损失

Abstract: Aiming at the scheduling problem for parallel batching machines with incompatible families and dynamic Lot arrivals in the cleaning-furnace zone of semiconductor production line， the Nested Partitions based on Differential Evolution（NPDE）algorithm was proposed. This algorithm made full use of advantages of global concurrent search of NPDE algorithm and local search capability of DE. To have a greater probability to track the most promising region of nested partitions, the DE was used to optimize the sample group from feasible region of nested partitions to reduce backtracking process. The data of simulation models and the actual production line were used to testify the effectiveness of the NPDE to minimize the Total Weighted Tardiness (TWT). Simulation results show that the propose algorithm was more effective than other heuristic algorithms.

Key words: semiconductor production line, parallel batching machine, optimization, scheduling, nested partitions, total weighted tardiness