Logo image
運用簡化群體演算法於產線平衡規劃問題-以半導體製程為例
Thesis

運用簡化群體演算法於產線平衡規劃問題-以半導體製程為例

丁奕安
Masters, 國立清華大學, 工業工程與工程管理學系
2013

Abstract

黃光顯影 半導體製造業 柔性運算 平行機台 機台負載平衡 Photolithography, Semiconductor Process Soft Computing Load Balancing Parallel Machines
Semiconductor Process plays an important role in competitive industries in Taiwan. Photolithography is the most complicated and key part in IC process flow, taking up about 40%~50% above all process. It is necessary to allocate mask, machine and wafer in the process flow appropriately. In order to reach efficiency and increase rate of equipment utilization, reducing cycle time(C/T) of wafer on each machine during photolithography is expected. It's capable of install coater, exposure and developer in series on the same machine to shorten C/T. However, there are still many methods to keep cutting down C/T .Except for adding parallel machine, running the existing parallel machine efficiently is also required. In other words, it makes every machine reach the best in the loading and reduce C/T on the restrictions of mask by transferring goods. It does works, and wafers are produced on time as expected. This research collected information of an IC manufactory in Science Park to set up a model of photolithography considering all the constraints on process. We found out a better wafer transfer dispatching solution through SSO algorithms and MM-SSO, and establishing a complete arithmetic logic applied to photolithography dispatching system. After testing and debugging, the logical operation can be applied to the follow-up study and demonstration in pratice. It's able to downsize and save time about 70% by algorithms automatic dispatching.

Metrics

1 Record Views

Details

Logo image