Prerequisite(s):
A statistical course
Text Book:
Montgomery, Douglas C., ‘Design and Analysis of Experiments’, Wiley, New
York, 5th Edition (2000)
Reference Books:
Box,
George E. P., Hunter, William G., Hunter, J. Stuart, ‘Statistics for
Experimenters’, Wiley, New York (1978)
Myers,
Raymound H. & Montgomery, Douglas, ‘Response Surface Methodology’, Wiley,
New York (1995)
Course
Description
There are techniques to design
an experiment and analyze the experimental data to reveal the effects of
factors namely process variables. This course is about these methods and
techniques, which provide a relationship between factors and response
variables (outputs). It emphasizes the connection between the experiment
and the model that the experimenter can develop from the results of the
experiment. As an introduction to the course, the fundamental concepts of
experimental design, such as randomization and blocking, comparison of
treatments, the analysis of variance along with simple graphical
techniques will be presented. Factorial and fractional factorial designs
with particular emphasis on the two-level design system will be
introduced. Fitting regression models (linear regression), Response
surface methods (RSM), which are the tools for process optimization trough
designed experiments, will be covered.
In many
industries, the effective use of statistical experimental design is the
key to higher yields, reduced variability, and better products. It is
believed that, this course can be very useful for students from all
science and engineering disciplines.
Course
Outline
I. Introduction
I1.
What is experimental design?
I2.
The role of experimental design
I3.
Basic statistics, Probability distributions
I4.
Normality checking
II. Comparing treatments
II1.
Significance tests (Hypothesis) and confidence intervals
for means and variances
II2.
Randomization and blocking with paired comparisons
II3.
Use of Analysis of Variance with a single factor
II4.
Randomization, blocking and Latin squares
III.
Factorial design experiments
III1. Weakness of Classical
one-variable at a time strategy
III2.
Introduction to Factorial Designs
III3. Two level Factorial
Designs
III4. Fractional Factorial
Designs
III5. Three-level and Mixed-level Factorial Designs
IV. Response Surface Methods
V1.
Simple modelling with Least Squares (Regression Models)
V2.
Central Composite Design
V3.
Box-Behnken Designs
Teaching Methods
Students
will be assigned homework every two weeks. Excel will be enough for the
early homework. Computer software including statistical functions is
required. There will be one midterm examination and a project. Students
are responsible for forming study groups of two and finding a study topic
for their term projects. A presentation of the projects will be given by
each group.
Students
can study with several design of experiment software that is available in
department’s computer laboratory:
Modde 7 (Umetrics)
Minitab 13
Design-Expert 6, trial version
(Stat-ease)
Grading
Homework assignments: 20
%
Midterm Exam: 30 %
Term Project: 40%
Presentation: 10%
Study
Material
DOE
Introduction (PDF file)
(Chapter 1 in Eriksson, L,
Johansson, E, Kettaneh-Wold, N, Wilkström, C & Wold, S, ‘Design of
Experiments: Principles and Applications’, Umetrics Academy, Umeå (2000))