一种人工智能AI搜索算法在集成电路生产制造动态调度的应用研究

Dynamic Scheduling Study of Integrated Circuit Manufactory Based on One AI Searching Arithmetic

集成电路制造的生产调度是典型的可重入生产调度系统。由于近年来集成电路制造工艺越来越复杂,制造成本及生产周期也在上升,如何合理的安排投产就变得日益重要。目前常用的调度方法是传统的基于规则的调度方法,如:先来先服务。这种方法缺乏系统性,往往造成较高的生产周期,造成客户不满。以生产周期最优为目的,尝试结合滚动窗口策略和某AI搜索算法(遗传算法),以期找到一种可行、有效、又具全局观的动态调度方法。

Scheduling of Integrated circuit manufactory is a classic re-entrant scheduling system. How to make it effectively become more and more important as the process becoming more and more complex, while its cost highly increased. Traditional scheduling methodology focused on rule setting, such as FIFO, which makes cycle time long and customers＇ dissatisfactory, is not systematic. This article aims at cycle time optimization, trying to combine genetic arithmetic, an AI searching arithmetic, with rolling policy, hoping to figure out a reasonable, effective and systematic dynamic scheduling method.