晶圓製造系統中以限制理論為基礎的多重式派工法則研究

Abstract

派工法則的研究是當前晶圓製造的重要課題，並且對系統績效影響甚大。在複雜的多產品晶圓混合生產中，一般派工法則不具備檢視晶圓在製品獲利能力，造成系統獲利無法提升，故侯氏[2001]藉著限制理論(Theory of Constraints, TOC)對系統資源消耗量的觀點，提出晶圓在製品現階段價值和獲利強度的估算法，發展出一系統化的限制理論基礎派工法則(Dispatch by theory of Constraint, DTC)。但是DTC忽略其他如交期、品質與在製品等因素，長期而言系統可能會受到罰金、商譽、品質不良和在製品囤積等因素影響系統最終獲利。且DTC在派工時如遇到在製品獲利強度相同而不足以決斷派工順序時，是以先到先做(First Come First Service, FCFS)法則僅考慮物件來到的先後順序派工，並無法根據現場的情況給予適當的派工。

本研究提出以交期方面表現優良之CR法則、在製品數量上表現優良之MCR法則和在產品品質上表現優良之FSMCT法則修正DTC，組合為一個以DTC為基礎之多重優先派工法則，使DTC提高對交期、在製品與品質的反應能力。 本研究以模擬模型，進行DTC與多重優先派工法則的績效評估。實驗結果發現多重優先派工法則可有效改善DTC於產品交期、在製品數量和產品品質等績效。

Job dispatch is an important issue in wafer fabrications. It affects system performance. However, few dispatching rules developed, in the past, focus on profit. Hou[2001] developed a TOC-based dispatching rule, named DTC (Dispatching by Theory of Constraints). Hou’s method based on the consumption of system resource can estimate the WIP cost and values. Based on the profit (the difference of values and cost) of the wafers, the job dispatching can then be assigned. Hence, DTC is superior to others from the profit viewpoint in a certain time-period. Can it meet other requirements such as due date, quality or/and minimum inventory? If it cannot, in the long run, DTC can never be a good dispatching rule from any view of points. Furthermore, in Hou’s study he uses FCFS to dispatch job when two jobs have the same profit. Instead of using wafer arriving time, it seems more effective to include due date, quality, or/and minimum inventory as the second and/or the third job dispatching criterion.

The main purpose of this study is to extend Hou’s simple-profit -based DTC to a multiple purpose dispatching method. Famous CR, MCR and FSMCT are included in the new method. CR is good in due date, MCR in low inventory, and FSMCT in quality. By simulations, several multiple purpose dispatching methods, by the combinations of DTC, CR, MCR and FSMCT, are performed. The simulation results indicate the best combination can effectively improve the performance of due date, quality and numbers of inventory.