Tchebycheff Metric 기반 가중평균제곱오차 최소화법을 활용한 다중반응표면 최적화

Abstract

Abstract Multiresponse optimization (MRO) seeks to find the setting of input variables, which optimizes the multipleresponses simultaneously. The approach of weighted mean squared error (WMSE) minimization for MRO imposesa different weight on the squared bias and variance, which are the two components of the mean squared error (MSE).To date, a weighted sum-based method has been proposed for WMSE minimization. On the other hand, this methodhas a limitation in that it cannot find the most preferred solution located in a nonconvex region in objective functionspace. This paper proposes a Tchebycheff metric-based method to overcome the limitations of the weighted sum-basedmethod. Key Words : Multiresponse Optimization, Weighted Mean Squared Error, Tchebycheff Metric, Weighted Sum 이 논문은 2012학년도 대구대학교 학술연구비 지원에 의한 논문임. * Corresponding Author : In-Junx Jeong (Daegu Univ.)Tel: +82-53-850-6275 email: ijjeong@daegu.ac.krReceived August 20, 2014 Revised October 13, 2014 Accepted January 8, 2015