Regression and Kriging metamodels with their experimental designs in simulation: A review

Research output: Contribution to journalLiterature reviewScientificpeer-review

Abstract

This article reviews the design and analysis of simulation experiments. It focusses on analysis via two types of metamodel (surrogate, emulator); namely, low-order polynomial regression, and Kriging (or Gaussian process). The metamodel type determines the design of the simulation experiment, which determines the input combinations of the simulation model. For example, a first-order polynomial regression metamodel should use a "resolution-III" design, whereas Kriging may use "Latin hypercube sampling". More generally, polynomials of first or second order may use resolution III, IV, V, or "central composite" designs. Before applying either regression or Kriging metamodeling, the many inputs of a realistic simulation model can be screened via "sequential bifurcation". Optimization of the simulated system may use either a sequence of low-order polynomials-known as 'response surface methodology" (RSM)- or Kriging models fitted through sequential designs- including "efficient global optimization" (EGO). Finally, "robust" optimization accounts for uncertainty in some simulation inputs.
Original languageEnglish
Pages (from-to)1-16
Number of pages16
JournalEuropean Journal of Operational Research
Volume256
Issue number1
DOIs
Publication statusPublished - Jan 2017

Fingerprint

Kriging
Metamodel
Experimental design
Design of experiments
Regression
Polynomial Regression
Polynomials
Simulation Experiment
Simulation
Simulation Model
Latin Hypercube Sampling
First-order
Sequential Design
Response Surface Methodology
Polynomial
Metamodeling
Robust Optimization
Gaussian Process
Bifurcation (mathematics)
Global Optimization

Keywords

  • robustness and sensitivity
  • metamodel
  • design
  • regression
  • kriging

Cite this

@article{bf7348fe35ff42ac998de047b5a600a1,
title = "Regression and Kriging metamodels with their experimental designs in simulation: A review",
abstract = "This article reviews the design and analysis of simulation experiments. It focusses on analysis via two types of metamodel (surrogate, emulator); namely, low-order polynomial regression, and Kriging (or Gaussian process). The metamodel type determines the design of the simulation experiment, which determines the input combinations of the simulation model. For example, a first-order polynomial regression metamodel should use a {"}resolution-III{"} design, whereas Kriging may use {"}Latin hypercube sampling{"}. More generally, polynomials of first or second order may use resolution III, IV, V, or {"}central composite{"} designs. Before applying either regression or Kriging metamodeling, the many inputs of a realistic simulation model can be screened via {"}sequential bifurcation{"}. Optimization of the simulated system may use either a sequence of low-order polynomials-known as 'response surface methodology{"} (RSM)- or Kriging models fitted through sequential designs- including {"}efficient global optimization{"} (EGO). Finally, {"}robust{"} optimization accounts for uncertainty in some simulation inputs.",
keywords = "robustness and sensitivity, metamodel, design, regression, kriging",
author = "J.P.C. Kleijnen",
year = "2017",
month = "1",
doi = "10.1016/j.ejor.2016.06.041",
language = "English",
volume = "256",
pages = "1--16",
journal = "European Journal of Operational Research",
issn = "0377-2217",
publisher = "Elsevier Science BV",
number = "1",

}

Regression and Kriging metamodels with their experimental designs in simulation: A review. / Kleijnen, J.P.C.

In: European Journal of Operational Research, Vol. 256, No. 1, 01.2017, p. 1-16.

Research output: Contribution to journalLiterature reviewScientificpeer-review

TY - JOUR

T1 - Regression and Kriging metamodels with their experimental designs in simulation: A review

AU - Kleijnen, J.P.C.

PY - 2017/1

Y1 - 2017/1

N2 - This article reviews the design and analysis of simulation experiments. It focusses on analysis via two types of metamodel (surrogate, emulator); namely, low-order polynomial regression, and Kriging (or Gaussian process). The metamodel type determines the design of the simulation experiment, which determines the input combinations of the simulation model. For example, a first-order polynomial regression metamodel should use a "resolution-III" design, whereas Kriging may use "Latin hypercube sampling". More generally, polynomials of first or second order may use resolution III, IV, V, or "central composite" designs. Before applying either regression or Kriging metamodeling, the many inputs of a realistic simulation model can be screened via "sequential bifurcation". Optimization of the simulated system may use either a sequence of low-order polynomials-known as 'response surface methodology" (RSM)- or Kriging models fitted through sequential designs- including "efficient global optimization" (EGO). Finally, "robust" optimization accounts for uncertainty in some simulation inputs.

AB - This article reviews the design and analysis of simulation experiments. It focusses on analysis via two types of metamodel (surrogate, emulator); namely, low-order polynomial regression, and Kriging (or Gaussian process). The metamodel type determines the design of the simulation experiment, which determines the input combinations of the simulation model. For example, a first-order polynomial regression metamodel should use a "resolution-III" design, whereas Kriging may use "Latin hypercube sampling". More generally, polynomials of first or second order may use resolution III, IV, V, or "central composite" designs. Before applying either regression or Kriging metamodeling, the many inputs of a realistic simulation model can be screened via "sequential bifurcation". Optimization of the simulated system may use either a sequence of low-order polynomials-known as 'response surface methodology" (RSM)- or Kriging models fitted through sequential designs- including "efficient global optimization" (EGO). Finally, "robust" optimization accounts for uncertainty in some simulation inputs.

KW - robustness and sensitivity

KW - metamodel

KW - design

KW - regression

KW - kriging

U2 - 10.1016/j.ejor.2016.06.041

DO - 10.1016/j.ejor.2016.06.041

M3 - Literature review

VL - 256

SP - 1

EP - 16

JO - European Journal of Operational Research

JF - European Journal of Operational Research

SN - 0377-2217

IS - 1

ER -