Abstract
The primary purpose of a company is profit making. There are many ways to improve the profitability of a company. For a semiconductor manufacturing company, yield improvement, rescheduling, and productivity enhancement can all lead to more profits. Much research has been done on scheduling in a single fab environment in recent years, but there is little research about multi-fabs orders allocation. The trend of the semiconductor manufacturing industry is “The big ones become bigger and the small ones become smaller”. Therefore many mergers are taking place. This causes some large companies to have more than one fab. How to allocate orders between multi-fabs and how to make sure the allocation can gain more profit become a very important topic for a company which has more than one fab.In the past, there was no good way to allocate customers’ orders between multi-fabs in a semiconductor manufacturing company. This situation is often found in companies whose core business is foundry because they must face unexpected orders from customers and they can only depend on past experience to do order allocation. And the allocation cannot make sure if it can bring maximum profit for the company. But they often do not take the profit of the whole company into account. If we want to get the maximum profit for the company, we must consider all fabs of the company and maximize the profit.This research proposes a linear programming model to solve this kind of problems. It considers the total waferout, yield, cost (including fixed and variable cost), price, special constraint of the fab, the capacity limit (photo moves of one fab) and the allocation ratio of one order. The objective is to maximize profit . A commercial software LINDO was applied to solve the LP problem. Real data collected from a semiconductor manufacturing company was tested for a system verification. The results show that the proposed model can effectively allocate the orders among different fabs.1.1研究背景與動機 11.2研究目的 41.3研究範圍與限制 71.4研究方法 8第二章 文獻探討 102.1生產管制活動簡介 102.2產能規劃 132.2.1粗略產能規劃 142.2.2產能需求規劃 172.2.3無限負荷細部產能規劃 202.2.4有限負荷細部產能規劃 202.2.5晶圓製程之產能負荷規劃方法 212.3訂單管理 242.4訂單分配 28第三章 產能與訂單分派問題 353.1問題描述與現況 353.2訂單處理程序 363.3決策程序建立 403.4 LP MODEL 413.5模式說明 533.5.1建構模式基本資料 533.5.2結果 563.5.3結果分析與討論 57第四章 模式驗證 614.1模式輸入資料與假設 614.2結果分析與討論 654.2.1CASE1 654.2.2CASE2 724.2.3CASE3 774.3 討論 82第五章 結論與未來展望 835.1結論 835.2未來研究方向 84參考文獻 86附錄A 模式1 90附錄B CASE1 95附錄C CASE2 100附錄D CASE3 105