Analysis of computer experiments using penalized likelihood in Gaussian kriging models

Runze Li, Agus Sudjianto

Research output: Contribution to specialist publicationArticle

77 Citations (Scopus)

Abstract

Kriging is a popular analysis approach for computer experiments for the purpose of creating a cheap-to-compute "meta-model" as a surrogate to a computationally expensive engineering simulation model. The maximum likelihood approach is used to estimate the parameters in the kriging model. However, the likelihood function near the optimum may be flat in some situations, which leads to maximum likelihood estimates for the parameters in the covariance matrix that have very large variance. To overcome this difficulty, a penalized likelihood approach is proposed for the kriging model. Both theoretical analysis and empirical experience using real world data suggest that the proposed method is particularly important in the context of a computationally intensive simulation model where the number of simulation runs must be kept small because collection of a large sample set is prohibitive. The proposed approach is applied to the reduction of piston slap, an unwanted engine noise due to piston secondary motion. Issues related to practical implementation of the proposed approach are discussed.

Original languageEnglish (US)
Pages111-120
Number of pages10
Volume47
No2
Specialist publicationTechnometrics
DOIs
StatePublished - May 1 2005

Fingerprint

Penalized Likelihood
Computer Experiments
Kriging
Simulation Model
Pistons
Experiments
Maximum likelihood
Likelihood Function
Metamodel
Maximum Likelihood Estimate
Covariance matrix
Maximum Likelihood
Theoretical Analysis
Engine
Model
Engineering
Motion
Estimate
Engines
Simulation

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Modeling and Simulation
  • Applied Mathematics

Cite this

@misc{d934e32923d64f0a9b9c56886a1583a7,
title = "Analysis of computer experiments using penalized likelihood in Gaussian kriging models",
abstract = "Kriging is a popular analysis approach for computer experiments for the purpose of creating a cheap-to-compute {"}meta-model{"} as a surrogate to a computationally expensive engineering simulation model. The maximum likelihood approach is used to estimate the parameters in the kriging model. However, the likelihood function near the optimum may be flat in some situations, which leads to maximum likelihood estimates for the parameters in the covariance matrix that have very large variance. To overcome this difficulty, a penalized likelihood approach is proposed for the kriging model. Both theoretical analysis and empirical experience using real world data suggest that the proposed method is particularly important in the context of a computationally intensive simulation model where the number of simulation runs must be kept small because collection of a large sample set is prohibitive. The proposed approach is applied to the reduction of piston slap, an unwanted engine noise due to piston secondary motion. Issues related to practical implementation of the proposed approach are discussed.",
author = "Runze Li and Agus Sudjianto",
year = "2005",
month = "5",
day = "1",
doi = "10.1198/004017004000000671",
language = "English (US)",
volume = "47",
pages = "111--120",
journal = "Technometrics",
issn = "0040-1706",
publisher = "American Statistical Association",

}

Analysis of computer experiments using penalized likelihood in Gaussian kriging models. / Li, Runze; Sudjianto, Agus.

In: Technometrics, Vol. 47, No. 2, 01.05.2005, p. 111-120.

Research output: Contribution to specialist publicationArticle

TY - GEN

T1 - Analysis of computer experiments using penalized likelihood in Gaussian kriging models

AU - Li, Runze

AU - Sudjianto, Agus

PY - 2005/5/1

Y1 - 2005/5/1

N2 - Kriging is a popular analysis approach for computer experiments for the purpose of creating a cheap-to-compute "meta-model" as a surrogate to a computationally expensive engineering simulation model. The maximum likelihood approach is used to estimate the parameters in the kriging model. However, the likelihood function near the optimum may be flat in some situations, which leads to maximum likelihood estimates for the parameters in the covariance matrix that have very large variance. To overcome this difficulty, a penalized likelihood approach is proposed for the kriging model. Both theoretical analysis and empirical experience using real world data suggest that the proposed method is particularly important in the context of a computationally intensive simulation model where the number of simulation runs must be kept small because collection of a large sample set is prohibitive. The proposed approach is applied to the reduction of piston slap, an unwanted engine noise due to piston secondary motion. Issues related to practical implementation of the proposed approach are discussed.

AB - Kriging is a popular analysis approach for computer experiments for the purpose of creating a cheap-to-compute "meta-model" as a surrogate to a computationally expensive engineering simulation model. The maximum likelihood approach is used to estimate the parameters in the kriging model. However, the likelihood function near the optimum may be flat in some situations, which leads to maximum likelihood estimates for the parameters in the covariance matrix that have very large variance. To overcome this difficulty, a penalized likelihood approach is proposed for the kriging model. Both theoretical analysis and empirical experience using real world data suggest that the proposed method is particularly important in the context of a computationally intensive simulation model where the number of simulation runs must be kept small because collection of a large sample set is prohibitive. The proposed approach is applied to the reduction of piston slap, an unwanted engine noise due to piston secondary motion. Issues related to practical implementation of the proposed approach are discussed.

UR - http://www.scopus.com/inward/record.url?scp=18944382184&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=18944382184&partnerID=8YFLogxK

U2 - 10.1198/004017004000000671

DO - 10.1198/004017004000000671

M3 - Article

VL - 47

SP - 111

EP - 120

JO - Technometrics

JF - Technometrics

SN - 0040-1706

ER -