Course Description
Covers state-of-the-art knowledge on selected topics such as factorial experiments, fractional factorials, incomplete block designs, orthogonal arrays, crossover designs, response surface methodology, mixture experiments, optimal design theory for linear and nonlinear models, and design construction techniques. Computer experiments and associated space-filling designs will be covered.
Athena Title
Adv in Experimental Desig
Prerequisite
STAT 8260 and STAT 6430
Semester Course Offered
Offered spring
Grading System
A - F (Traditional)
Course Objectives
Students will learn fundamental concepts and advanced methods for designing an experiment and analyzing data from the experiment. Students will learn special treatment structures, including factorial treatment structures. Inference methods for assessing differences among treatments or for specified comparisons of interest are discussed in a general framework that uses linear models approach. They will learn about the theory of optimal design, both for linear and nonlinear models, and how this can help in selecting efficient designs. They will learn how to make inferences based on the methods learned in the course using statistical software. They will also explore the use of software for finding efficient designs. Furthermore, they will learn techniques for the construction of a variety of designs. Students will learn about complex computer models and associated space-filling designs, with applications in different scientific and engineering contexts.
Topical Outline
The topics covered in this course focus on different methods for collecting data in designed experiments and computer experiments and on the analysis of such data. Designs considered include incomplete block designs, fractional factorial designs, orthogonal arrays, Latin square designs, and other row-column designs. Complex computer models with engineering and scientific applications will be covered. Space-filling designs and Kriging models will be discussed. Design construction along with numerical optimization techniques will also be considered. Commonly used design optimality criteria will be introduced and discussed. These will be applied in a variety of situations, and software for finding optimal or efficient designs will also be discussed. Additional topics that may be covered depending on time and interest include split-plot designs, crossover designs, designs and models for response surface methodology, designs and models for mixture experiments, choice experiments, and conjoint analysis.