Kriging metamodeling for simulation

W.C.M. Beers, van

Research output: ThesisPhd Thesis 4 Research NOT TU/e / Graduation NOT TU/e)

Abstract

Many scientific disciplines use mathematical models to describe complicated real systems. Often, analytical methods are inadequate, so simulation is applied. This thesis focuses on computer intensive simulation experiments in Operations Research/Management Science. For such experiments it is necessary to apply interpolation. In this thesis, Kriging interpolation for random simulation is proposed and a novel type of Kriging - called Detrended Kriging - is developed. Kriging turns out to give better predictions in random simulation than classic low-order polynomial regression. Kriging is not sensitive to variance heterogeneity: i.e. Kriging is a robust method. Moreover, the thesis develops a novel method to select experimental designs for expensive simulation. This method is sequential, and accounts for the specific input/output function implied by the underlying simulation model. For deterministic simulation the designs are constructed through cross-validation and jackknifing, whereas for random simulation the customization is achieved through bootstrapping. The novel method simulates relatively more input combinations in the interesting parts of the input/output function, and gives better predictions than traditional Latin Hypercube Sample designs with prefixed sample sizes.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • Universiteit vanTilburg
Supervisors/Advisors
  • Kleijnen, J.P.C., Promotor, External person
Award date28 Oct 2005
Place of PublicationTilburg
Publisher
Publication statusPublished - 2005

Fingerprint

Interpolation
Management science
Operations research
Design of experiments
Experiments
Polynomials
Mathematical models

Cite this

Beers, van, W. C. M. (2005). Kriging metamodeling for simulation. Tilburg: Universiteit vanTilburg.
Beers, van, W.C.M.. / Kriging metamodeling for simulation. Tilburg : Universiteit vanTilburg, 2005. 105 p.
@phdthesis{5e8065ffef17403b96cc5257c5b5d2f9,
title = "Kriging metamodeling for simulation",
abstract = "Many scientific disciplines use mathematical models to describe complicated real systems. Often, analytical methods are inadequate, so simulation is applied. This thesis focuses on computer intensive simulation experiments in Operations Research/Management Science. For such experiments it is necessary to apply interpolation. In this thesis, Kriging interpolation for random simulation is proposed and a novel type of Kriging - called Detrended Kriging - is developed. Kriging turns out to give better predictions in random simulation than classic low-order polynomial regression. Kriging is not sensitive to variance heterogeneity: i.e. Kriging is a robust method. Moreover, the thesis develops a novel method to select experimental designs for expensive simulation. This method is sequential, and accounts for the specific input/output function implied by the underlying simulation model. For deterministic simulation the designs are constructed through cross-validation and jackknifing, whereas for random simulation the customization is achieved through bootstrapping. The novel method simulates relatively more input combinations in the interesting parts of the input/output function, and gives better predictions than traditional Latin Hypercube Sample designs with prefixed sample sizes.",
author = "{Beers, van}, W.C.M.",
year = "2005",
language = "English",
publisher = "Universiteit vanTilburg",
school = "Universiteit vanTilburg",

}

Beers, van, WCM 2005, 'Kriging metamodeling for simulation', Doctor of Philosophy, Universiteit vanTilburg, Tilburg.

Kriging metamodeling for simulation. / Beers, van, W.C.M.

Tilburg : Universiteit vanTilburg, 2005. 105 p.

Research output: ThesisPhd Thesis 4 Research NOT TU/e / Graduation NOT TU/e)

TY - THES

T1 - Kriging metamodeling for simulation

AU - Beers, van, W.C.M.

PY - 2005

Y1 - 2005

N2 - Many scientific disciplines use mathematical models to describe complicated real systems. Often, analytical methods are inadequate, so simulation is applied. This thesis focuses on computer intensive simulation experiments in Operations Research/Management Science. For such experiments it is necessary to apply interpolation. In this thesis, Kriging interpolation for random simulation is proposed and a novel type of Kriging - called Detrended Kriging - is developed. Kriging turns out to give better predictions in random simulation than classic low-order polynomial regression. Kriging is not sensitive to variance heterogeneity: i.e. Kriging is a robust method. Moreover, the thesis develops a novel method to select experimental designs for expensive simulation. This method is sequential, and accounts for the specific input/output function implied by the underlying simulation model. For deterministic simulation the designs are constructed through cross-validation and jackknifing, whereas for random simulation the customization is achieved through bootstrapping. The novel method simulates relatively more input combinations in the interesting parts of the input/output function, and gives better predictions than traditional Latin Hypercube Sample designs with prefixed sample sizes.

AB - Many scientific disciplines use mathematical models to describe complicated real systems. Often, analytical methods are inadequate, so simulation is applied. This thesis focuses on computer intensive simulation experiments in Operations Research/Management Science. For such experiments it is necessary to apply interpolation. In this thesis, Kriging interpolation for random simulation is proposed and a novel type of Kriging - called Detrended Kriging - is developed. Kriging turns out to give better predictions in random simulation than classic low-order polynomial regression. Kriging is not sensitive to variance heterogeneity: i.e. Kriging is a robust method. Moreover, the thesis develops a novel method to select experimental designs for expensive simulation. This method is sequential, and accounts for the specific input/output function implied by the underlying simulation model. For deterministic simulation the designs are constructed through cross-validation and jackknifing, whereas for random simulation the customization is achieved through bootstrapping. The novel method simulates relatively more input combinations in the interesting parts of the input/output function, and gives better predictions than traditional Latin Hypercube Sample designs with prefixed sample sizes.

UR - http://repository.uvt.nl/id/ir-uvt-nl:oai:wo.uvt.nl:172768

M3 - Phd Thesis 4 Research NOT TU/e / Graduation NOT TU/e)

PB - Universiteit vanTilburg

CY - Tilburg

ER -

Beers, van WCM. Kriging metamodeling for simulation. Tilburg: Universiteit vanTilburg, 2005. 105 p.