臺灣博碩士論文加值系統

生產晶圓時，減少產品的單位成本是重要的一環。為此，準確地預測產品的單位成本是必須的，其中包括兩項工作，成本的計算和未來成本的預估。大部分的文獻均著重於成本的計算，而較少探討未來成本的預估；本研究提出混合模糊線性迴歸與倒傳遞網路的方法來預估單位晶粒的有效成本(effective cost per die)，由於良率學習的不確定性，使得預測單位晶粒的有效成本十分困難。在實務上，長期的有效成本，根據成本持續下降的管理哲學，常被評估為一線性迴歸方程式，這是不正確的。與傳統的線性迴歸法相較，模糊線性迴歸法有許多優點。從模糊線性迴歸法所求得的預測值，我們利用倒傳遞網路予以解模糊化以得到一明確值，在此領域為一新的方法。為了評估此方法的有效性，我們套用實際的數據。實驗結果證明模糊線性迴歸合併倒傳遞網路法優於現有的一些方法，讓預測晶粒的有效成本更加的正確和精確。

Reducing the unit cost for every product type is a critical task to a wafer fabrication plant. For this purpose, accurately estimating the unit cost is the prerequisite, which involves two sub-tasks: costing and future cost estimation. Most references in this field were focused on costing and seldom investigated the estimation of future unit cost. A hybrid fuzzy linear regression (FLR) and back propagation network (BPN) approach is proposed to estimate the effective cost per die for every product type in a wafer fabrication plant, which is difficult because of the uncertainty in yield learning. In practical situations the long-term effective cost per die is usually estimated with a linear regression (LR) equation, according to the “continuous cost down” philosophy, which is not accurate. Conversely, the proposed FLR approach has many advantages. Afterwards, a BPN is constructed to defuzzify the fuzzy estimated cost and to derive a representative value, which is also novel theoretically. For evaluating the effectiveness of the proposed methodology, some real data were collected. Experimental results showed that the hybrid FLR-BPN approach was superior to some existing approaches in estimation accuracy and precision.

誌謝 i
中文摘要 ii
Abstract iii
目錄 iv
圖目錄 vi
表目錄 vii
第一章 緒論 1
1.1 研究背景 1
1.2 研究動機 2
1.3 研究目的 3
1.4 研究方法 5
1.5 論文架構及研究流程 6
第二章 文獻探討 7
2.1 半導體成本的預測 7
2.2 傳統預測成本的方法 8
2.3 良率對成本的影響 9
2.4 模糊線性迴歸模型 13
2.5 類神經網路 15
2.5.1 類神經網路簡介 15
2.5.2倒傳遞網路 17
第三章 研究方法 19
3.1 模糊線性規劃合併倒傳遞網路法 19
3.2 建構模糊線性規劃模型 20
3.3 建構考慮專家意見的模糊線性迴歸模型 24
3.4應用模糊交集 27
3.5應用倒傳遞網路解模糊化 28
第四章 實例運算與分析 29
4.1 實例驗證 29
4.2 使用傳統的良率學習模型 30
4.3 使用模糊線性迴歸模型 32
4.3.1 模糊線性迴歸模型 32
4.3.2 考慮專家意見的模糊線性迴歸模型 37
4.4 模糊類神經方法 39
4.4.1 模糊交集 40
4.4.2 倒傳遞網路 42
第五章 結論與未來研究方向 46
參考文獻 48

1.http://www.moeaidb.gov.tw/external/view/tw/index.html (2010)，「經濟部工業局」
2.http://cdnet.stpi.org.tw/techroom.htm (2010)，「科技產業資訊室」
3.http://cdnet.stpi.org.tw/techroom/policy/2009/policy_09_106.htm (2010)，「科技產業資訊室」
4.http://www.moea.gov.tw/ (2010)，「經濟部，工業生產變動概況」
5.http://2k3dmz2.moea.gov.tw/gnweb/News/wFrmNews_Detail.aspx?id=161 (2010)，「經濟部統計處」
6.http://www.eettaiwan.com/ (2010)，「電子工程專輯」
7.http://www.pvtaiwan.com/chinese/overview/semi.shtml (2010)，「國際半導體設備材料產業協會」
8.唐麗英，謝昆霖, 「積體電路良率管理之良率模式構建-類神經網路之應用」，南華大學學術期刊，資訊管理研究，第3卷，民國92年7月
9.黃仁鵬、郭英哲、楊怡和，「應用倒傳遞類神經網路於網路教學之自動化學習效果評量」， 2004 國際學術研討會 資訊科技與實務
10.葉懷仁，「類神經網路應用於引擎聲訊之辨識」，國立中山大學碩士論文，民國92年1月
11.邱挺晏，「消費分期市場評等評分模型建置的方向」，民國94年6月
12.Carnes, R. (1991), “Long-term cost of ownership: beyond purchase price,” Proceedings of 1991 IEEE/SEMI International Semiconductor Science Symposium, 39-43
13.Chen, T., and Wang, M. J. J. (1999), “A fuzzy set approach for yield learning modeling in wafer manufacturing,” IEEE Transactions on Semiconductor Manufacturing, 12(2), 252-258.
14.Chen, T., and Lin, Y. C. (2008), “A fuzzy-neural system incorporating unequally important expert opinions for semiconductor yield forecasting,” International Journal of Uncertainty, Fuzziness, and Knowledge-based Systems, 16(1), 35-58.
15.Chen, T. (2009), “A fuzzy-neural knowledge-based system for job completion time prediction and internal due date assignment in a wafer fabrication plant,” International Journal of Systems Science, in press.
16.Eric Hsiu-Chun Wang, (1999) ” An Integrated Framework for Yield Learning in Semiconductor Manufacturing”, A DISSERTATION SUBMIITTED TO THE DEPARTMENT OF ENGINEERING-ECONOMIIC SYSTEMS
17.Gruber, H. (1994), Learning and Strategic Product Innovation: Theory and Evidence for the Semiconductor Industry, Elsevier.
18.Min Liu,, Yean Yng Ling., (2003) “Using fuzzy neural network approach to estimate contractors’ markup” Building and Environment, 38, 1303 -1308
19.Peters, G. (1994), “Fuzzy linear regression with fuzzy intervals,” Fuzzy Sets and Systems, 63, 45-55
20.Sal Agnihothri*, Ji Soo Lee, Jeongseob Kim (2000) “Lot sizing with random yields and tardiness costs” Computers & Operations Research 27 ,437-459
21.Shih-Lin Hung, Ching-Yun Kao, Ming-Fu Shieh., (2001)” APPLICATION OF BACK-PROPAGATION NEURAL NETWORKS IN NONLINEAR HYSTERETIC STRUCTURAL DYNAMIC ANALYSIS” Journal of the Chinese Institute of Civil and Hydraulic Engineering, Vol. 13, No. 3, 567–583
22.Toly Chen.,(2003), “A fuzzy back propagation network for output time prediction in a wafer fab” Applied Soft Computing, 2(3), 211–222.
23.Tanaka, H., and Watada, J. (1988), “Possibilistic linear systems and their application to the linear regression model,” Fuzzy Sets and Systems, 272, 275-289.
24.Pfitzner, L., Benesch, N., ?鉍hsner, R., Schmidt, C., Schneider, C., Tschaftary, T., Trunk, R., and Dudenhausen, H.M. (2001), “Cost reduction strategies for wafer expenditure,” Microelectronics Engineering, 56, 61-71.
25.Wood, S. C. (1997), “Cost and cycle time performance of fabs based on integrated single-wafer processing,” IEEE Transactions on Semiconductor Manufacturing, 10(1), 98-111.
26.Watada, J., and Tanaka, H., and Shimomura, T (1986), “Identification of learning curve based on possibilistic concepts”, Applications of Fuzzy Set Theory in Human Factors, The Netherlands: Elsevier Science Publishers B.V.
27.Yuantu Huang., Tom D. Gedeon,, Patrick M. Wong., (2001) “An integrated neural-fuzzy-genetic-algorithm using hyper-surface membership functions to predict permeability in petroleum reservoirs” Engineering Applications of Artificial Intelligence 14,15-21