Application-driven Sequential Designs for Simulation Experiments: Kriging Metamodeling

J.P.C. Kleijnen, W.C.M. van Beers

Research output: Working paperDiscussion paperOther research output

310 Downloads (Pure)

Abstract

This paper proposes a novel method to select an experimental design for interpolation in simulation.Though the paper focuses on Kriging in deterministic simulation, the method also applies to other types of metamodels (besides Kriging), and to stochastic simulation.The paper focuses on simulations that require much computer time, so it is important to select a design with a small number of observations.The proposed method is therefore sequential.The novelty of the method is that it accounts for the specific input/output function of the particular simulation model at hand; i.e., the method is application-driven or customized.This customization is achieved through cross-validation and jackknifing.The new method is tested through two academic applications, which demonstrate that the method indeed gives better results than a design with a prefixed sample size.
Original languageEnglish
Place of PublicationTilburg
PublisherOperations research
Number of pages24
Volume2003-33
Publication statusPublished - 2003

Publication series

NameCentER Discussion Paper
Volume2003-33

Fingerprint

Design of experiments
Interpolation
Experiments

Keywords

  • experimental design
  • simulation
  • interpolation
  • sampling
  • sensitivity analysis
  • metamodels

Cite this

Kleijnen, J. P. C., & van Beers, W. C. M. (2003). Application-driven Sequential Designs for Simulation Experiments: Kriging Metamodeling. (CentER Discussion Paper; Vol. 2003-33). Tilburg: Operations research.
Kleijnen, J.P.C. ; van Beers, W.C.M. / Application-driven Sequential Designs for Simulation Experiments : Kriging Metamodeling. Tilburg : Operations research, 2003. (CentER Discussion Paper).
@techreport{1af5ac49a5f34b8a901cb094a45ea518,
title = "Application-driven Sequential Designs for Simulation Experiments: Kriging Metamodeling",
abstract = "This paper proposes a novel method to select an experimental design for interpolation in simulation.Though the paper focuses on Kriging in deterministic simulation, the method also applies to other types of metamodels (besides Kriging), and to stochastic simulation.The paper focuses on simulations that require much computer time, so it is important to select a design with a small number of observations.The proposed method is therefore sequential.The novelty of the method is that it accounts for the specific input/output function of the particular simulation model at hand; i.e., the method is application-driven or customized.This customization is achieved through cross-validation and jackknifing.The new method is tested through two academic applications, which demonstrate that the method indeed gives better results than a design with a prefixed sample size.",
keywords = "experimental design, simulation, interpolation, sampling, sensitivity analysis, metamodels",
author = "J.P.C. Kleijnen and {van Beers}, W.C.M.",
note = "Pagination: 24",
year = "2003",
language = "English",
volume = "2003-33",
series = "CentER Discussion Paper",
publisher = "Operations research",
type = "WorkingPaper",
institution = "Operations research",

}

Kleijnen, JPC & van Beers, WCM 2003 'Application-driven Sequential Designs for Simulation Experiments: Kriging Metamodeling' CentER Discussion Paper, vol. 2003-33, Operations research, Tilburg.

Application-driven Sequential Designs for Simulation Experiments : Kriging Metamodeling. / Kleijnen, J.P.C.; van Beers, W.C.M.

Tilburg : Operations research, 2003. (CentER Discussion Paper; Vol. 2003-33).

Research output: Working paperDiscussion paperOther research output

TY - UNPB

T1 - Application-driven Sequential Designs for Simulation Experiments

T2 - Kriging Metamodeling

AU - Kleijnen, J.P.C.

AU - van Beers, W.C.M.

N1 - Pagination: 24

PY - 2003

Y1 - 2003

N2 - This paper proposes a novel method to select an experimental design for interpolation in simulation.Though the paper focuses on Kriging in deterministic simulation, the method also applies to other types of metamodels (besides Kriging), and to stochastic simulation.The paper focuses on simulations that require much computer time, so it is important to select a design with a small number of observations.The proposed method is therefore sequential.The novelty of the method is that it accounts for the specific input/output function of the particular simulation model at hand; i.e., the method is application-driven or customized.This customization is achieved through cross-validation and jackknifing.The new method is tested through two academic applications, which demonstrate that the method indeed gives better results than a design with a prefixed sample size.

AB - This paper proposes a novel method to select an experimental design for interpolation in simulation.Though the paper focuses on Kriging in deterministic simulation, the method also applies to other types of metamodels (besides Kriging), and to stochastic simulation.The paper focuses on simulations that require much computer time, so it is important to select a design with a small number of observations.The proposed method is therefore sequential.The novelty of the method is that it accounts for the specific input/output function of the particular simulation model at hand; i.e., the method is application-driven or customized.This customization is achieved through cross-validation and jackknifing.The new method is tested through two academic applications, which demonstrate that the method indeed gives better results than a design with a prefixed sample size.

KW - experimental design

KW - simulation

KW - interpolation

KW - sampling

KW - sensitivity analysis

KW - metamodels

M3 - Discussion paper

VL - 2003-33

T3 - CentER Discussion Paper

BT - Application-driven Sequential Designs for Simulation Experiments

PB - Operations research

CY - Tilburg

ER -

Kleijnen JPC, van Beers WCM. Application-driven Sequential Designs for Simulation Experiments: Kriging Metamodeling. Tilburg: Operations research. 2003. (CentER Discussion Paper).