2003 Quality & Productivity Research Conference

Conference Theme:Ê Data + Statistics = Competitive Advantage

Tutorial: "Experimenting with Mixtures", by Prof. John A. Cornell

Many industrial processes involve the blending of ingredients to form end products. Experimenting with these processes consists of varying the proportions (or percentages) of the individual ingredients in an effort to see if measured quality characteristics or properties of the end product change from one blend to the next. For example, a property of interest of stainless steel, which is a mixture of iron, nickel, copper, and chromium, is its tensile strength. Changing the relative proportions of iron, nickel, and copper, while holding the proportion of chromium fixed will certainly affect the tensile strength. Changing the percentages of orange, pineapple, and grapefruit juices, respectively, from 30%-35%-35% to 40%-50%-10% would certainly change the flavor of the three-juice beverage.
The workshop will begin by addressing the questions, ãWhat are mixture experiments?ä and ãHow do mixture experiments differ from ordinary experiments such as factorial experiments?ä Topics to be covered are,

* Designs and models for exploring the simplex-shaped factor space.
* The analysis of mixture data.
* Additional constraints on the component proportions.
* Designs (and software) and data analysis techniques for constrained mixture regions.
* Model forms other than the Scheffe-type models.
* The inclusion of process variables and/or varying the total amount of the mixture.
* Identifying optimum formulations.
* Many worked examples from the initial design to inferences from the data from real experiments will be provided.

The workshop will benefit anyone wanting to learn about statistical techniques for designing mixture experiments, analyzing the resulting data, and optimizing product formulations. Pre- requisites are an understanding of elementary statistical concepts and hypothesis testing. Some previous exposure to non-mixture experimental design (factorial, fractional factorial and response surface designs) and least squares regression would be very helpful.

Introducing the Instructor:

Dr. John A. Cornell is Professor and Statistician with the Department of Statistics and Agricultural Experiment Station at the University of Florida. He has served as a consultant to industry and in the agricultural sciences in the area of mixture experiments for the past 35 years. He has been an active researcher during this time, with over 140 publications in the fields of experimental design, mixture experiments, and other areas. He is the author of the text
Experiments with Mixtures, 3rd edition, John Wiley & Sons, Inc., and co-author of Response Surfaces, Designs and Analyses, 2nd edition, Marcel Dekker, with A.I. Khuri. Professor Cornell is a past editor of the Journal of Quality Technology, a Fellow of the American Statistical Association and of the American Society for Quality (ASQ) and was awarded The Shewart Medal from ASQ for the year 2000.

Registration Form