AN APPROACH FOR DETERMINING THE COST MATRIX OF MULTIVARIATE QUALITY LOSS FUNCTION


Dolgun L. E., Eriskin L., Ozkan G., KÖKSAL G.

INTERNATIONAL JOURNAL OF INDUSTRIAL ENGINEERING-THEORY APPLICATIONS AND PRACTICE, vol.27, no.6, pp.879-905, 2020 (SCI-Expanded) identifier

  • Publication Type: Article / Article
  • Volume: 27 Issue: 6
  • Publication Date: 2020
  • Doi Number: 10.23055/ijietap.2020.27.6.6265
  • Journal Name: INTERNATIONAL JOURNAL OF INDUSTRIAL ENGINEERING-THEORY APPLICATIONS AND PRACTICE
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, Compendex
  • Page Numbers: pp.879-905
  • Keywords: multivariate quality loss function, cost matrix, multiple response optimization, multi-response design parameter optimization, robust product and process design, preferential dependency, OPTIMAL TOLERANCE DESIGN, ROBUST DESIGN, OPTIMIZATION, PRODUCTS, UNCERTAINTY
  • TED University Affiliated: No

Abstract

Multivariate quality loss functions are commonly used in product and process design parameter optimization, which involves simultaneous consideration of multiple responses in the determination of the levels of design parameters that provide high-quality performance. These functions are also used in statistical tolerancing and quality improvement decision making. This study investigates the bivariate loss function in terns of its ability to represent different values or preferences a decision-maker may attribute to different settings of the responses. Then, an interactive and evolutionary method for estimating the cost matrix parameters of the multivariate quality loss function is proposed. The method is applied to a hypothetical bivariate loss function parameter estimation problem as well as a real trivariate case associated with a honing operation. It is shown on the hypothetical problem that the method converges to the true underlying loss function after a few iterations even when the information provided by the decision-maker contains certain degrees of errors. The convergence is also shown for different variance-covariance structures.