【文章內容簡介】 ，因為我們看不到兩個或更多因子同時改變時的效果。這是說因子 A在因子 B處于高水平時和處于低水平時對輸出的影響可能是不一樣的。 OFAT Six Sigma Green Belt Training Design of Experiments (DoE) Read the case material provided and prepare the following as teams: 閱讀所提供的案例材料并分組回答下面問題： ? What was different about the designed experiment that Alex ran as pared to the OFAT experiment? Alex運行的經過設計的實驗與 OFAT實驗有什么不同？ ? Is there now a pointer to an area where yields may be higher? 現在是否有指示出更高產出率的方向？ ? What new inference space should Alex explore? Alex應該探求的新推論空間是什么？ Case Study – Yield Improvement Sections 4 amp。 5 案例分析 – 改善產出率之第四、五節(jié) Six Sigma Green Belt Training Design of Experiments (DoE) ? Full Factorial Designs (Factorial designs) test every possible bination of factors over the inference （因子實驗）檢驗推論空間范圍內的所有可能的組合 ? These designs are of particular interest because they provide a great deal of information. 這些設計極其重要因為它們 提供了大量的信息 。 ? The amount of resources necessary to run full factorial designs can be exorbitant if the number of factors is large. Full factorials are generally run when the experimenter has a high degree of confidence that the factors in the study have an influence over the ，運行全因子實驗所必需的資源數量是非常大的。一般實驗者在有高度的信心認為所研究的因子對輸出有影響時才運行全因子實驗。 Full Factorials 全因子實驗 因子數量 (2水平 ) 運行次數 2 22 = 4 3 23 = 8 4 24 = 16 5 25 = 32 6 26 = 64 2n Factor columns (2n): 因子列 (2n): Exponent。 n = of factors to be tested 指數 。 n = 因子數 Base。 2 = the of levels to be tested for each factor 基數 。 2 = 每個因子的水平數 Six Sigma Green Belt Training Design of Experiments (DoE) DOE Design and Results: DOE設計和結果： Which Factors appear to be important? 哪一個參數看起來重要些 ? How should the important factors be set? 這些參數應該如何設定 ? Is there an Interaction between the variables? 因子之間是否存在交互作用 ? Full Factorial – Yield Improvement DOE 案例分析 – 改善產出率之 DOE Time Temp 時間 溫度 Yield A B AB 產出率 70 () 145 () + 56 130 (+) 145 () 69 70 () 165 (+) 82 130 (+) 165 (+) + 58 Six Sigma Green Belt Training Design of Experiments (DoE) ? The first consideration before applying any statistical analysis technique is whether the results of the DOE are of any practical 統(tǒng)計工具之前要考慮其結果是否有實用價值。 ? Did the response variable change? Did it change the desired amount?響應變量會隨之變化嗎 ? 變化量是否足夠大 ? ? If the response variable did not change substantially across the factor treatment binations, it may be that: 如果相應變量變化不夠明顯，可能是因為 : 1. Factor levels were not set far enough apart (didn’t go bold enough).因子水平設置不夠大 (不夠大膽 ) 2. The selected factors do not affect the response 影響 3. The measurement system is not ? Practical 實用 ? Graphical 圖示 ? Analytical 解析 Six Sigma Rules of Analysis: 六西格瑪分析的規(guī)則 ? Practical實用 Six Sigma Green Belt Training Design of Experiments (DoE) ? Another