- 學位論文
遺傳演算法為基的滾動式時窗排程法求解動態黃光區派工問題
A Time Window Rolling- and GA-Based Method for the Dynamic Dispatching Problem in Photolithography Area
摘要
黃光微影製程是晶圓製造重要的瓶頸製程。在製程良率的考量下，黃光區派工問題具有多樣且複雜的製程限制，常以人工方式進行黃光區派工。本研究經與業界多次訪談後，建立符合實務的求解模式，綜合考量緊急晶圓批優先權、機台關鍵層負載量平衡、關鍵層綁機台限制、機台製程能力限制、以及實務上會加入排程的光罩試曝和擋片作業等。在此模式下，本研究提出一以遺傳演算法為基的時窗滾動式黃光區機台排程優化法，求解實際的黃光區機台派工問題。此時窗滾動為基的作法係由當下時間往後延伸一時段形成時窗，考量時窗內的晶圓批排程需求。排程對象是目前已在黃光區及時窗內預期到達的晶圓批。隨著時間的遞移，此時窗往前滾動進行遺傳演化重排程。停留時間加權值最小化、非製程加工的額外時間最小化、最晚完工時間最小化、和機台執行關鍵層負載不平均量最小化是本研究定義的四個優化子目標，引導排程朝著提昇機台利用率、增加產出量、降低晶圓批週期時間、及減少額外作業時間浪費的結果演化。為驗證本求解模式的成效及實用性，本研究並實作一套「遺傳演算之時窗滾動式黃光區排程系統」，以具物件導向特性的黃光區模擬實驗平台，讀入實際黃光區運作的晶圓到達歷史資料，模擬時窗滾動觸發排程系統進行排程派工，設定晶圓批層級加工的機台及排定加工時程。本研究並定義平均機台利用率、完成量、晶圓批平均生產週期時間、額外作業時間等四個效能指標，以評斷排程方法的成效。實驗數據顯示本研究提出的排程方法優於人工派工及簡單的先到先派工法。此外實驗結果也顯示在適當的時窗時程設定下，整個派工成效有顯著的提昇。
並列摘要
The operation management of photolithography area is the most important yet difficult task in semiconductor manufacturing. Due to numerous constraints and operation goals imposed, manual decision and action are adopted in scheduling and dispatching the jobs for the machines in this area. The operation requirements and constraints considered in this paper include lot-priority, machine load balance for critical layers, machine bounded by critical layers, heterogeneous processing capability, reticle pilot run, dummy wafer test, etc. Under these considerations, this paper presents a time window rolling- and GA-based scheduling system to assign and schedule both the arrived and on-the-way wafer lots to the photolithography machines. Time window is defined when the scheduler is triggered by extending a time period to round in the wafer lots on the way to the photo area within this window. The scheduled jobs are then executed accordingly until the next scheduling event is triggered; and thus the widow rolls. The presented model proposed four minimization sub-goals, weighted cycle time, extra time, makespan, and critical layer loading unbalanced value, to conduct the scheduling results to enhancing the machine utilization and throughput while reducing cycle time and nonproductive time. To verify the proposed GA model, a prototype system namely “Photolithography Area Simulation System,” was developed implementing the proposed scheduling method and a simulation platform for the photolithography area. In addition, four performance indices are proposed, tool utilization, throughput, cycle time, and extra operation time, for evaluating the scheduling methods. Numerical tests were conducted using historical operation data from a Taiwanese DRAM manufacturing factory. Production activities are simulated by the historical data and executed by the scheduling results from the proposed model. Results showed that the proposed method outperformed the manual one and a simple FCFS scheduling method. In particular, the performance improvement reached a peak subject to a time window width specification.