頁籤選單縮合
題 名 | Neural-Taguchi Method for Robust Design Analysis=新的強健設計分析工具一類神經田口法 |
---|---|
作 者 | 王國禎; 蔡志成; 曾柏昌; 陳琮祺; | 書刊名 | 中國機械工程學刊 |
卷 期 | 19:2 1998.04[民87.04] |
頁 次 | 頁223-230 |
分類號 | 440.8 |
關鍵詞 | 類神經網路田口法; Neural network; Taguchi method; Orthogonal array; Sensitivity analysis; Parameters design and optimization; |
語 文 | 英文(English) |
中文摘要 | Neural-Taguchi method, a new engineering analysis method that combines the neural-network and the Taguchi techniques for better improvement of engineering design is proposed in this paper. In this new Neural-Taguchi mothod, or thogonal array, which requires only partial experiments is executed first. Each row in the orthogonal array along with its relative responses ( mean η and variance σ 2 ) form a set of training pattern to the neural network. Implicit system model that allows precise prediction and easy optimizing is constructed in terms of a multilayer feedforward neural network. Since an implicit model is obtained, interpolation, extrapolation, optimization and prediction can be easy to conduct. A die cast component analysis example is also used to illustrated the advantage and performance of the proposed Neural-Taguchi method. |
英文摘要 | 本研究乃是結合類神經網路與田口實驗法而發展出一能有效改善工程設計分析程序之 方法。在此一新的類神經網路田口法中,直交表被用來作為實驗設計的工具,期能在最少實 驗次數下獲取足夠之系統輸入出關係。直交表每一列之元素加上其相對應之系統輸出平均值 (η)與變異數(σ),構成一組類神經網路訓練樣本。訓練完成之類神經網路即是一能提 供精確內插、外插、預測與最佳化搜尋之系統模式。而此一新的類神經網路田口法之功能亦 透過一模具鑄件之實例加以驗證。 |
本系統中英文摘要資訊取自各篇刊載內容。