Date of Defense

12-9-1994

Date of Graduation

Spring 4-22-1995

Department

Statistics

First Advisor

John Crowell, Mathematics and Statistics

Second Advisor

Frank Severance, Electrical Engineering

Abstract

Mixtures of ingredients occur often in both everyday and industrial settings. When forming a mixture, the concern is often how much of each ingredient should be added to create a mixture which is the "best" in some respect. If a numerical value, called the response, can be assigned to how "good" a mixture is, this problem becomes a natural candidate for statistical analysis. This paper discusses several of the issues which arise when analyzing data from mixture experiments and how a computer can help. After a brief overview of the basics of mixture experiments, the formation of a response equation will be discussed. Following, a method is presented for finding the best combination of ingredients, called components, under some special circumstances. Throughout, the macro capabilities of the statistical package Minitab are used to implement the proposed algorithms.

Access Setting

Honors Thesis-Campus Only

Download

Share

COinS