A simplex search method for experimental optimization with multiple objectives

MARTINEZ, ERNESTO

Computer-aided Chemical Engineering

Elsevier

Lugar: Amsterdam; Año: 2006; p. 377 - 382

Experimental optimization with multiple responses in the face of noise and outliers is a key issue in product/process development and to guarantee product end-use properties. A novel concept of maximimizing the concordance of desirability functions of all responses concerned is presented. The Kendall's coefficient of concordance W borrowed from nonparametric statistics is used to provide a statistical characterization of optimality in a multiobjective setting. A multi-directional simplex method is proposed in the concordance function framework. Simplex reflection, expansion and contraction operations are based on ranking vertices according to their corresponding values of W.