Monotonicity-Preserving Bootstrapped Kriging Metamodels for Expensive Simulations

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

Research output: Working paperDiscussion paperOther research output

221 Downloads (Pure)

Abstract

Kriging (Gaussian process, spatial correlation) metamodels approximate the Input/Output (I/O) functions implied by the underlying simulation models; such metamodels serve sensitivity analysis and optimization, especially for computationally expensive simulations. In practice, simulation analysts often know that the I/O function is monotonic. To obtain a Kriging metamodel that preserves this known shape, this article uses bootstrapping (or resampling). Parametric bootstrapping assuming normality may be used in deterministic simulation, but this article focuses on stochastic simulation (including discrete-event simulation) using distribution-free bootstrapping. In stochastic simulation, the analysts should simulate each input combination several times to obtain a more reliable average output per input combination. Nevertheless, this average still shows sampling variation, so the Kriging metamodel does not need to interpolate the average outputs. Bootstrapping provides a simple method for computing a noninterpolating Kriging model. This method may use standard Kriging software, such as the free Matlab toolbox called DACE. The method is illustrated through the M/M/1 simulation model with as outputs either the estimated mean or the estimated 90% quantile; both outputs are monotonic functions of the traffic rate, and have nonnormal distributions. The empirical results demonstrate that monotonicity-preserving bootstrapped Kriging may give higher probability of covering the true simulation output, without lengthening the confidence interval.
Original languageEnglish
Place of PublicationTilburg
PublisherOperations research
Number of pages26
Volume2009-75
Publication statusPublished - 2009

Publication series

NameCentER Discussion Paper
Volume2009-75

Fingerprint

Discrete event simulation
Sensitivity analysis
Sampling

Keywords

  • Queues

Cite this

Kleijnen, J. P. C., & van Beers, W. C. M. (2009). Monotonicity-Preserving Bootstrapped Kriging Metamodels for Expensive Simulations. (CentER Discussion Paper; Vol. 2009-75). Tilburg: Operations research.
Kleijnen, Jack P.C. ; van Beers, W.C.M. / Monotonicity-Preserving Bootstrapped Kriging Metamodels for Expensive Simulations. Tilburg : Operations research, 2009. (CentER Discussion Paper).
@techreport{59d5c29b25a34af9921fbf3a28d93a38,
title = "Monotonicity-Preserving Bootstrapped Kriging Metamodels for Expensive Simulations",
abstract = "Kriging (Gaussian process, spatial correlation) metamodels approximate the Input/Output (I/O) functions implied by the underlying simulation models; such metamodels serve sensitivity analysis and optimization, especially for computationally expensive simulations. In practice, simulation analysts often know that the I/O function is monotonic. To obtain a Kriging metamodel that preserves this known shape, this article uses bootstrapping (or resampling). Parametric bootstrapping assuming normality may be used in deterministic simulation, but this article focuses on stochastic simulation (including discrete-event simulation) using distribution-free bootstrapping. In stochastic simulation, the analysts should simulate each input combination several times to obtain a more reliable average output per input combination. Nevertheless, this average still shows sampling variation, so the Kriging metamodel does not need to interpolate the average outputs. Bootstrapping provides a simple method for computing a noninterpolating Kriging model. This method may use standard Kriging software, such as the free Matlab toolbox called DACE. The method is illustrated through the M/M/1 simulation model with as outputs either the estimated mean or the estimated 90{\%} quantile; both outputs are monotonic functions of the traffic rate, and have nonnormal distributions. The empirical results demonstrate that monotonicity-preserving bootstrapped Kriging may give higher probability of covering the true simulation output, without lengthening the confidence interval.",
keywords = "Queues",
author = "Kleijnen, {Jack P.C.} and {van Beers}, W.C.M.",
note = "Pagination: 26",
year = "2009",
language = "English",
volume = "2009-75",
series = "CentER Discussion Paper",
publisher = "Operations research",
type = "WorkingPaper",
institution = "Operations research",

}

Kleijnen, JPC & van Beers, WCM 2009 'Monotonicity-Preserving Bootstrapped Kriging Metamodels for Expensive Simulations' CentER Discussion Paper, vol. 2009-75, Operations research, Tilburg.

Monotonicity-Preserving Bootstrapped Kriging Metamodels for Expensive Simulations. / Kleijnen, Jack P.C.; van Beers, W.C.M.

Tilburg : Operations research, 2009. (CentER Discussion Paper; Vol. 2009-75).

Research output: Working paperDiscussion paperOther research output

TY - UNPB

T1 - Monotonicity-Preserving Bootstrapped Kriging Metamodels for Expensive Simulations

AU - Kleijnen, Jack P.C.

AU - van Beers, W.C.M.

N1 - Pagination: 26

PY - 2009

Y1 - 2009

N2 - Kriging (Gaussian process, spatial correlation) metamodels approximate the Input/Output (I/O) functions implied by the underlying simulation models; such metamodels serve sensitivity analysis and optimization, especially for computationally expensive simulations. In practice, simulation analysts often know that the I/O function is monotonic. To obtain a Kriging metamodel that preserves this known shape, this article uses bootstrapping (or resampling). Parametric bootstrapping assuming normality may be used in deterministic simulation, but this article focuses on stochastic simulation (including discrete-event simulation) using distribution-free bootstrapping. In stochastic simulation, the analysts should simulate each input combination several times to obtain a more reliable average output per input combination. Nevertheless, this average still shows sampling variation, so the Kriging metamodel does not need to interpolate the average outputs. Bootstrapping provides a simple method for computing a noninterpolating Kriging model. This method may use standard Kriging software, such as the free Matlab toolbox called DACE. The method is illustrated through the M/M/1 simulation model with as outputs either the estimated mean or the estimated 90% quantile; both outputs are monotonic functions of the traffic rate, and have nonnormal distributions. The empirical results demonstrate that monotonicity-preserving bootstrapped Kriging may give higher probability of covering the true simulation output, without lengthening the confidence interval.

AB - Kriging (Gaussian process, spatial correlation) metamodels approximate the Input/Output (I/O) functions implied by the underlying simulation models; such metamodels serve sensitivity analysis and optimization, especially for computationally expensive simulations. In practice, simulation analysts often know that the I/O function is monotonic. To obtain a Kriging metamodel that preserves this known shape, this article uses bootstrapping (or resampling). Parametric bootstrapping assuming normality may be used in deterministic simulation, but this article focuses on stochastic simulation (including discrete-event simulation) using distribution-free bootstrapping. In stochastic simulation, the analysts should simulate each input combination several times to obtain a more reliable average output per input combination. Nevertheless, this average still shows sampling variation, so the Kriging metamodel does not need to interpolate the average outputs. Bootstrapping provides a simple method for computing a noninterpolating Kriging model. This method may use standard Kriging software, such as the free Matlab toolbox called DACE. The method is illustrated through the M/M/1 simulation model with as outputs either the estimated mean or the estimated 90% quantile; both outputs are monotonic functions of the traffic rate, and have nonnormal distributions. The empirical results demonstrate that monotonicity-preserving bootstrapped Kriging may give higher probability of covering the true simulation output, without lengthening the confidence interval.

KW - Queues

M3 - Discussion paper

VL - 2009-75

T3 - CentER Discussion Paper

BT - Monotonicity-Preserving Bootstrapped Kriging Metamodels for Expensive Simulations

PB - Operations research

CY - Tilburg

ER -

Kleijnen JPC, van Beers WCM. Monotonicity-Preserving Bootstrapped Kriging Metamodels for Expensive Simulations. Tilburg: Operations research. 2009. (CentER Discussion Paper).