透過您的圖書館登入
IP:3.139.81.58
  • 學位論文

最小偏誤設計之建構研究

Research on construction of minimum aberration design

指導教授 : 王丕承
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


實驗設計主要的目的是希望透過實驗的方式,檢測出主要影響產品特性值或反應值的因子與交互作用。由於資源有限,因此希望能夠藉由少數的處理組合,取得最多的資訊來處理問題。若實驗者將因子任意指派至直交表上,那麼在資料分析的過程中,可能會發生效應互相混淆的情形,因此主要的課題便是該如何選擇最佳的設計來安排實驗。通常會使用最小偏誤設計的準則來判斷設 計的優劣,其目的在於減少效應混淆的情形。故將因子安排至最小偏誤設計中,可讓實驗者更容易了解各種效應對於反應值所產生的影響。 而Chen, Sun & Wu (1993)就以最小偏誤設計的準則為主,在不同的因子個數與處理組合個數之下,提供了一系列最小偏誤設計的目錄,以供實驗者在指派因子時作為參考使用。而本研究主要目的是希望不需透過查詢目錄的方式,可以快速且直接建構出最小偏誤設計來使用。因此選取直交表中某部份的行作為主要的基本架構,利用刪減因子建構行或者增加其餘因子建構行的方式來建構出各種最小偏誤設計。只要了解兩種方法的原則,便可簡單的建構出最小偏誤設計。

並列摘要


The purpose of Design of Experiments is to find out factors and interactions that affect response by experiment. If experimenters arrange factors arbitrarily on the orthogonal array, it may be occur that the effects are aliased. So a key question is how to choose a fraction of the orthogonal array to arrange factors. The experimenters always use the minimum aberration design that can be estimate the most main effects. According to the minimum aberration criterion, Chen, Sun & Wu (1993) propose an algorithm for constructing complete catalogue of fractional factorial designs. The issue of this studying is to construct minimum aberration design easily and directly without referring CSW's catalogue. Therefore, we choose fraction of the orthogonal array to be basic construction, and propose two methods to construct the minimum aberration design.

參考文獻


[1] Box, G. E. P. and Hunter, J. S. (1961), “The 2n-p Fractional Factorial Designs Part I.” Technometrics, 3, 311−352.
[2] Chen, J., Sun, D. X. and Wu, C. F. J. (1993), “A Catalogue of Two-Level and Three-Level Fractional Factorial Designs With Small Runs.” International Statistical Review, 61, 131−145.
[4] Fries, A. & Hunter, W. G. (1980), “Minimum Aberration 2k-p designs.” Technometrics, 22, 601−608.
[6] Li, H. and Lin, D.K.J. (2003), “Optimal Foldover Plans for Fractional Factorial Designs.” Technometrics, 45, 142−149.
[7] Li, H. and Mee, R. W. (2002), “Better Foldover Fractions for Resolution Ⅲ 2k-p Designs.” Technometrics, 44, 278−283.

延伸閱讀