中文摘要：

为了调度具有大规模、多重入、混合型生产等特征的半导体晶圆制造系统,提出一种将基于问题分解的方法与蚁群算法相结合的新的调度方法.通过基于时间与基于系统结构的复合分解过程,将大规模复杂调度问题分解为若干子问题,在相邻子问题间设置重叠操作以保证调度的一致性.为调度各子问题,提出一种分类蚁群算法以处理不同类型的机器.仿真实验在一个虚拟的晶圆制造系统上进行,结果表明,该方法对调度结果有显著的提升.

英文摘要：

According to unique features of semiconductor wafer fabrication system （SWFS）, such as large scale, multiple re-entrant, and multi-product manufacturing, this paper proposed a new scheduling method by combining decomposition-based method and ant colony optimization （ACO） algorithm. A large and complicated scheduling problem is decomposed into several subproblems by the compound decomposition procedure according to the features of time-sequence and system structures. Overlapping operations are set between neighbor subproblems to ensure the consistency. To schedule the subproblems, the classified- ACO algorithm was proposed to deal with the different type machines. The experiments were executed on a virtual SWFS fab simulation platform, and the results show the method can significantly improve the performance of schedule.