晶圓製造廠訂單分類模式之構建

Abstract

近幾年來，半導體廠商數量大增，生產型態轉為供過於求，使得市場競爭更加激烈， 因此半導體廠商在無法預知客戶的下單模式，而又必須提供客戶最佳服務，滿足客戶需求 以掌握客源的要求下，緊急訂單的規劃管理問題也逐漸凸顯出來。本文之研究目的即在建 立一套緊急訂單分類模式，希望以一合理的方法來決定系統中應該存在的最適緊急訂單比 例及緊急訂單等級數，以協助管理者作接單時的參考依據。 本研究主要分兩個階段，首先在模擬分析階段即是以模擬的方式探討與緊急訂單相關 的因子對系統造成的影響，再以統計手法分析得知緊急訂單比例之不同及緊急訂單優先權 等級數之多寡對於系統中一般訂單之生產週期時間、一般訂單及緊急訂單之生產週期時間 變異、一般訂單的產出量皆有顯著性之影響，以這些受影響之項目當作績效因子，做為吾 人考量緊急訂單價格及緊急訂單承接問題之衡量依據。 下一階段緊急訂單分類階段即是由利潤的觀點出發，由管理者決定出所能承受一般訂 單延遲的最大容許量限制，再以統計機率分配求出可能延遲之一般訂單數量及天數，用以 求算處罰成本。然後計算承接緊急訂單所能多加之利潤，將之扣除延遲之處罰成本，得到 系統整體利潤。透過此種利潤分析的觀點，探討在該系統環境下，系統應存在之最適緊急 訂單比例及緊急訂單優先權等級數，以追求利潤極大化。 吾人之研究，可提供管理決策者在承接緊急訂單時有效的資訊，使之瞭解是否該承接 此訂單，以及承接後系統利潤的變化情形。同時吾人之研究亦可提供顧客下緊急訂單時參 考的資訊，使之瞭解在其要求之生產週期時間下，廠商的訂價及所能保證的達交率，作為 顧客在下單時權衡的依據。

The numbers of semiconductor manufacturing factories become more and more in recent years. This situation makes the supply of products be more than the demand, and causes the market competition to be heavy. Because of the difficulty of accurate order forecasting and the need of customer service improving, the management an dplanning of rush orders becomes important. The goal of this thesis is to construct a model of rush order classification, for deciding the proper rush ordrs' proportion and the number of priority levels existed int he system. This information is useful for managers to decide whether or not to accept the rush orders. This research has two parts. First, use the simulation to find out all the factors that have effects on the system performance because of the introducing or rush orders. Then the statistical analysis finds out that the cycle time and throughput amount of normal orders, the cycle time variation of normal and rush orders are significantly impacted by rush orders. These significantly impacted items are used as factors for measuring the rush order's price an dfor deciding the acceptance of rush orders. Second, use the probability distribution of cycle time to find out the delay orders and the corresponding delay days that are out of the system allowable tolerance. Then calculate the profit gained by accepting rush orders and the penalty incurred because of order delay. The system profits under different situations are compared so as to choose the proper rush orders' proportion and priority numbers. The result of this research provides much information of system performance about accepting rush orders. It is useful for managers and customers. The competitive advantage of the semiconductor factory thus is nhanced.