實驗設計與統計
胡子陵
立德管理學院資環所副教授
Leader University
design and analysis of experiments
1
實驗設計與統計課程綱要及進度









Week1：Statistics (Review)
Week2：Simple Comparative Experiments
Week3：Experiments with a Single Factor：The Analysis of
Variance(1)
Week4：Experiments with a Single Factor：The Analysis of
Variance(2)
Week5：Introduction to Factorial Designs (1)
Week6：Introduction to Factorial Designs (2)
Week7：The 2K Factorial Design(1)
Week8：The 2K Factorial Design(2)
Week9：-----Mid-Term Report--------
design and analysis of experiments
2
實驗設計與統計課程綱要及進度









Week10：Blocking and Confounding in the 2k Factorial Design(1)
Week11：Blocking and Confounding in the 2k Factorial Design(2)
Week12：Two-Level Fractional Factorial Design(1)
Week13：Two-Level Fractional Factorial Design(2)
Week14：Fitting Regression Models(1)
Week15：Fitting Regression Models(2)
Week16：Response Surface Methods and Other Approaches to
Process Optimization(1)
Week17：Response Surface Methods and Other Approaches to
Process Optimization (2)
Week18：----Final Examination----------
design and analysis of experiments
3
實驗設計與統計
Part 1 – 前言
Chapter 1
課程目的
 歷史回顧
 一些基本原理(原則)和術語
 實驗的策略
 規畫、進行和分析實驗的方針

design and analysis of experiments
4
實驗設計介紹


An experiment is a test or a series of tests
Experiments are used widely in the
engineering world
 Process
characterization & optimization
 Evaluation of material properties
 Product design & development
 Component & system tolerance determination

“All experiments are designed experiments,
some are poorly designed, some are welldesigned”
design and analysis of experiments
5
解釋名詞






Response variable
Explanatory variable
Treatment
Lurking variable
Confounded
Statistical significance：我們的結論有統計顯著
性，即證據或結果強到很少會光靠機遇(chance)
而發生。
design and analysis of experiments
6
工程上之實驗要求





Reduce time to
design/develop new
products & processes
Improve performance of
existing processes
Improve reliability and
performance of products
Achieve product & process
robustness
Evaluation of materials,
design alternatives, setting
component & system
tolerances, etc.
design and analysis of experiments
7
實驗設計歷史發展的基本原則

Randomization
 Running
the trials in an experiment in random order
 Notion of balancing out effects of “lurking” variables

Replication
 Sample
size (improving precision of effect estimation,
estimation of error or background noise)
 Replication versus repeat measurements?

Blocking
 Dealing
with nuisance factors
design and analysis of experiments
8
The meaning of Blocking
Blocking(區集)：一組實驗個體，這些個
體在實驗之前，就被認為在會影響反應的
某些地方很近似；與抽樣中的分層樣本具
有相同類似的功用。
 Blocking design：將個體隨機分派到各處
理的此一步驟，是在每個Blocking裡個別
執行的。

design and analysis of experiments
9
實驗策略

“Best-guess” experiments
 Having
a lot of technical or theoretical knowledge
 More successful than you might suspect, but there
are disadvantages…

One-factor-at-a-time (OFAT) experiments
associated with the “scientific” or
“engineering” method
 Devastated by interaction, also very inefficient
 Sometimes

Statistically designed experiments
 Based
on Fisher’s factorial concept
design and analysis of experiments
10
因子設計


In a factorial experiment,
all possible
combinations of factor
levels are tested
The golf experiment:








Type of driver
Type of ball
Walking vs. riding
Type of beverage
Time of round
Weather
Type of golf spike
Etc, etc, etc…
design and analysis of experiments
11
One-factor-at-a-time設計
Using the oversized driver, balata ball,
walking, and drinking water levels of the
four factors as the baseline.
Optimal combination: regular-sized driver,
riding, and drinking water
design and analysis of experiments
12
因子設計
design and analysis of experiments
13
多因子之因子設計
design and analysis of experiments
14
多因子之因子設計－部份因子
design and analysis of experiments
15
實驗設計指南
1.
2.
3.
4.
5.
6.
7.
Recognition of & statement of problem
Choice of factors, levels, and ranges
Selection of the response variable(s)
Choice of design
Conducting the experiment
Statistical analysis
Drawing conclusions, recommendations
design and analysis of experiments
16
實驗設計歷史發展的四個時期




The agricultural origins, 1918 – 1940s
 R. A. Fisher & his co-workers
 Profound impact on agricultural science
 Factorial designs, ANOVA (analysis of variance)
The first industrial era, 1951 – late 1970s
 Box & Wilson, response surfaces
 Applications in the chemical & process industries
The second industrial era, late 1970s – 1990
 Quality improvement initiatives in many companies
 Taguchi and robust parameter design, process
robustness
The modern era, beginning circa 1990
design and analysis of experiments
17
實驗中使用統計技術




Get statistical thinking involved early
Your non-statistical knowledge is crucial to
success
Pre-experimental planning (steps 1-3) vital
Think and experiment sequentially (use the
KISS principle)
design and analysis of experiments
18