晶圓廠短期動態機台調整機制

Abstract

由於晶圓代工廠的機台具有多功特性，可執行多種不同的加工作業。藉由調整某些機台的加工參數，來增加瓶頸機台群的產能，使系統產出增加。將此問題稱為動態產能調整問題。 本研究針對上述問題，提出短期動態機台調整機制，以最大化瓶頸機台群的產出量為目標，並滿足「高機台利用率之機台群，其產出量不得低於未調整機台前」的限制，來尋找最佳的調機方案。 本研究所提的方法，其基本的構想在決定選擇替代機台進行調整功能的優先順序。所謂替代機台是指，機台可調整其功能參數，以執行瓶頸機台群的加工作業。從替代機台的產能狀況、支援瓶頸機台的程度、及調整機台功能所需的設置時間三方面，來決定替代機台的優先順序，選擇順位高的替代機台進行功能調整，並藉由模擬來計算新機台組合的產出。重複上述步驟，當瓶頸機台群產出不再增加，則確認調機方案。最後，以一實例來說明並驗證所提之方法。

Some machines in an IC foundry are multifunctional; that is, they can be reset up in order to perform different operations. One way to quickly increase the throughput is by resetting up some machines in order to increase the capacity of the bottleneck machine group. This is known as the dynamic capacity adjustment problem. An algorithm for solving the problem is proposed in this research. The algorithm aims to maximize the throughput of the bottleneck machine group while keeping not decreasing the local throughput (moves) of some other highly utilized machine groups. A machine is called an alternative machine if it can be reset up and becomes a member in the bottleneck machine group. The basic idea of the algorithm is to determine the priorities of these alternative machines, and the alternative machine with the highest priority is selected to move into the bottleneck machine group. The expected throughput of the new machine mix configuration is then calculated through simulation for the justification of the capacity adjustment plan. The above procedure is continuously performed until the objective function cannot be improved any further. This algorithm has been implemented and shown satisfactory results in a testing case.