Multiresponse Optimization: An Exploratory Study with Gaussian Process Model  
Author Nuno Costa

 

Co-Author(s) João Lourenço

 

Abstract To optimize multiple quality characteristics of process and product is a frequent problem in industry, and Ordinary Least Squares is the most popular technique for fitting models to responses. However, recent advances in computational power, the need to better understand and improve systems (processes and products) as well as to solve new and increasingly complex real-life problems have impelled researchers to develop and use a large variety of regression techniques. This article explores the usefulness of Gaussian Process model in the Response Surface Methodology framework, using a classical example from the literature. Results provide evidence that Gaussian Process model is an alternative to popular Ordinary Least Squares modeling technique.

 

Keywords Covariance, Experiments, Modeling, OLS, Regression, RSM
   
    Article #:  20151
 
Proceedings of the 20th ISSAT International Conference on Reliability and Quality in Design
August 7-9, 2014 - Seattle, Washington, U.S.A.