透過您的圖書館登入
IP:18.191.13.255
  • 學位論文

馬氏田口系統-兩階段最佳化、類神經網路演算法在動態環境資料探勘之應用

Modeling a dynamic design system using the Mahalanobis Taguchi system - two steps optimal based neural network

指導教授 : 蘇哲平 許總欣

摘要


本研究係以製造業產品/檢驗系統設計為例,來進行關鍵功能屬性與參數屬性萃取之資料挖掘分析之探討,以期以新的產品/檢驗系統模式來進行相關活動。 由於,傳統產品/檢驗系統之設計,大都建基於產品功能/檢驗屬性之設計,並未對關鍵功能/檢驗參數屬性來進行分析與萃取,以達穩健設計為目標。另外,當產品/檢驗系統模式一旦建立後,在動態環境下,此一模式是否適用的問題。因此,本研究回顧了相關產品設計、資料挖掘等相關文獻進行探討後,提出一個動態產品/檢驗系統參數設計的演算法:馬氏田口系統-兩階段最佳化以及類神經網路演算法-來解決動態環境資料挖掘系統設計模式建立問題。其中,馬氏田口系統演算法是田口先生所發展出來,解決模式建立問題;兩階段最佳化是田口所提出來,解決不同時段之模式建構問題;類神經網路於1982年被霍普菲爾(Hopfield)提出來處理輸入/輸出之間的問題,後漸漸亦用於動態環境下之輸入/輸出問題之模式建構。因此,此演算法係整合了馬氏田口系統、田口之兩階段最佳化、以及類神經網路等演算法,來進行動態環境下產品/系統參數選擇與設計之演算法。其中,馬氏田口系統演算法有別於傳統演算法之參數設計,而是對關鍵參數萃取;兩階段最佳化與類神經網路則是驗證此動態模式是否適當。透過研究驗證,馬氏田口系統可以有效地應用於產品/系統參數設計與選擇等之模式建立;而兩階段最佳化與類神經網路演算法可以成功且有效地運用在動態環境下,資料之探勘與模式建立上。 最後,藉由信賴水準與個案驗證之實驗發現,馬氏田口系統-兩階段最佳化以及類神經網路演算法,可以很容易且有效地解決動態模式建立問題。

並列摘要


This work presents a novel algorithm, the MTS-TSO based Neural Network (NN) algorithm, which combines the Mahalanobis Taguchi System (MTS) with the Two-Step Optimal (TSO) method for parameter selections which are adjusted under a dynamic environment for product parameter design. The utility of the algorithm is assessed in two dimensions-the MTS shows how individual product parameter dimensions are selected; and, the TSO-NN links parameters selection decisions across two different times and it can be used to focus on dynamic system design (DSD) and to identify product architecture dimensions that are critical for a dynamic design system strategy. The MTS which can easily solve product parameter design problems and shows it’s computationally efficient in the previous works. Additionally, the TSO algorithm is a simple and efficient means of constructing a dynamic design system, which is verified by the neural network algorithm from this work, and the neural network is already successfully applied in dynamic system of the past studies. Based on the main aims and verifies of this work, we conclude that the MTS-TSO based neural network algorithm can be applied successfully to dynamic environments for solving product design problems.

參考文獻


Ambrozic, T., and Turk, G. (2003), Prediction of subsidence due to underground mining byarti.cialne ural networks, Computers & Geosciences, Vol. 29, pp. 627–637.
Balakrisnan, S. N., and Weil, R. D. (1996), Neurocontrol: A Literature Survey, Mathematical and Computer Modelling, Vol. 23, Issue: 1-2, January, pp. 101-117.
Bendoly, E. (2003), Theory and support for process frameworks of knowledge discovery and data mining from ERP systems, Information & Management, Vol. 40. pp. 639–647.
Benner, M., Linnemann, A.R.,Jongen, W.M.F., and Folstar, P. (2003), Quality Function Deployment (QFD)—can it be used to develop food products?, Food Quality and Preference, Vol. 14, Issue: 4, June, pp. 327-339.
Boothroyd, G., Dewhurst, P., Knight, W.A. (2002), Product Design for Manufacture and Assembly Marcel Dekker, New York.

被引用紀錄


陳淑娟(2004)。心智繪圖融入國小低年級寫作教學之行動研究〔碩士論文,國立臺灣師範大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0021-2004200709120032
黃雅卿(2006)。國中地理科心智繪圖教學方案對學生創造力、學業成就表現之成效研究〔碩士論文,國立臺灣師範大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0021-2304200714090901
黃雅卿(2006)。國中地理科心智繪圖教學方案對學生創造力、學業成就表現、學習歷程之成效研究〔碩士論文,國立臺灣師範大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0021-0712200716091410
磨文慈(2009)。心智圖法教學策略對亞斯柏格症學生之成效研究〔碩士論文,國立臺灣師範大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0021-1610201315154564
林佩玲(2009)。寓言教學導引國中學生思維發展之研究〔碩士論文,國立臺灣師範大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0021-1610201315151696

延伸閱讀