摘要
在流体网络模型的基础上建立了以减小平均在制品水平目标的半导体制造系统的调度模型.这些模型通常很难求解, 但它们是可分解的, 即可以把整个调度周期上的调度问题分解为一系列子区间上的调度子问题. 在半导体生产环境下, 子问题可以进一步简化. 把上述结果当作调度规则直接应用于基于仿真的调度方法, 可以克服流体网络模型的近似性, 同时, 由于充分利用了系统全局的状态信息, 提高了仿真调度的优化能力.
We present a scheduling model with objective of optimizing the average WIP (Work In Process) of semiconductor manufacturing systems based on the fluid network model. It is usually very difficult to solve these models, but they are decomposable. This means the complex scheduling problems on the whole scheduling horizon can be decomposed into a series subproblems on individual subareas of time. The subproblems can be further simplified in the context of semiconductor manufacturing. Above results can be applied as a scheduling rule to simulation-based scheduling, which is used to improve the approximation of the fluid network model. Because this new scheduling method uses the global state information in decision, the optimization ability and performance of simulation-based scheduling are improved.
出处
《系统仿真学报》
EI
CAS
CSCD
2001年第2期133-135,138,共4页
Journal of System Simulation
基金
西安交通大学科研基金
