晶圓製造廠主生產排程之構建

Abstract

大部分的存貨式生產的半導體晶圓製造廠規劃生產計畫及主生產排程時， 並未充分掌握生產線機台產能、負荷等相關資訊，到了細部排程（Wafer Start Schedule）階段，才檢測生產線的產能負荷狀況。因此，若發現線 上產能無法滿足細部排程時，往往又為了達到生產計畫，仍舊按所規劃的 細部排程下料。結果，造成生產線在製品數量增加、生產週期時間加長， 產出卻沒有相對地增加，徒增生產線管理複雜度。因此本文在主生產排程 規劃階段，即結合生產線機台產能、負荷等相關資訊，發展一主生產排程 演算法檢核生產計畫是否合理可行，並在考量生產線平衡(Line Balance) 及充分使用瓶頸機器群產能的前提下，對不合理的生產計畫提供因應的調 整方案，擬定一合理(capacity feasible)的生產計畫。最後選擇以 Visual Foxpro3.0a資料庫程式語言作為本演算法程式撰寫工具。經由一 實例應用，並以機器群產能利用率來評估本演算法，結果在規劃幅度內本 演算法可提高機器群產能利用率(負荷/可用產能)並能將瓶頸機器群的負 荷平準化(Leveling)且充分應用瓶頸機器群的產能以因應市場對產品需求 量的變動。 When making production plan and master production schedule(MPS) , most of the make to stock wafer fabrication facilities fail to utilize the information of capacity and load in shop floor . It is not until the phase of wafer start schedule can the capacity and load be verified . In order to fulfill production plan , wafers are released according to the wafer start schedule though the capacity can not satisfy it . Consequently , it results in increase of WIP and cycle time , but the throughput does not improve accordingly . It just causes the complexity of shop floor management .Therefore , this research develops an MPS algorithm which integrates the information of capacity and load to verify the feasible production plan . In addition , this research provides the adjustment scheme to tackle the infeasible production plan by joint consideration of line balance and full utilization of bottleneck . Thus , a production plan and MPS which are capacity feasible can be made.Visual Foxpro 3.0a , a database language is selected to program this algorithm . In the case study , capacity utilization of machine group is used to evaluate this algorithm . This result reveals that this algorithm can enhance capacity utilization of machine group in the planning horizon . Moreover , leveling of bottleneck can be made and capacity of bottleneck can fully utilized to meet the variation of market demand .