System development and application of Taylor Kriging metamodeling

Abstract

PurposeThe purpose of this paper is to develop an application software interpolation system based on Taylor Kriging (TK) metamodeling, and apply the developed software system to addressing some engineering interpolation problems.Design/methodology/approachTK is a novel Kriging model where Taylor expansion is used to identify the base functions of drift function in Kriging. The paper explains the methodology of TK, illustrates the development of software, and reports the results of two case studies by comparing TK with several regression methods.FindingsTK has the advantage of interpolation accuracy, and the developed Kriging software system is useful and can be conveniently manipulated by users.Practical implicationsThe developed software system can benefit practical engineering applications that need accurate interpolations under limited observations.Originality/valueThis paper develops an application software interpolation system based on a novel TK metamodel, and the practical engineering applications show that it can provide accurate interpolations under limited observations.

Analytical target cascading using ensemble of surrogates for engineering design problems

“…Surrogate models have been extensively adopted in MDO to alleviate the computational burden in computational simulations (Gao et al, 2012). Most researchers are familiar with polynomial response surface (PRS; Chen et al, 2010), some other researchers use more sophisticated surrogate models such as Kriging (KRG; Liu et al, 2011Liu et al, , 2013Afonso et al, 2011), radial basis function (RBF; Bernal and Kindelan, 2010;Zheng et al, 2013a) and so on. Sellar et al (1996) introduced PRS into CSSO to reduce the computational cost.…”