Computational improvements to estimating Kriging metamodel parameters

Research output: Contribution to journalArticle

37 Citations (Scopus)

Abstract

The details of a method to reduce the computational burden experienced while estimating the optimal model parameters for a Kriging model are presented. A Kriging model is a type of surrogate model that can be used to create a response surface based a set of observations of a computationally expensive system design analysis. This Kriging model can then be used as a computationally efficient surrogate to the original model, providing the opportunity for the rapid exploration of the resulting tradespace. The Kriging model can provide a more complex response surface than the more traditional linear regression response surface through the introduction of a few terms to quantify the spatial correlation of the observations. Implementation details and enhancements to gradient-based methods to estimate the model parameters are presented. It concludes with a comparison of these enhancements to using maximum likelihood estimation to estimate Kriging model parameters and their potential reduction in computational burden. These enhancements include the development of the analytic gradient and Hessian for the log-likelihood equation of a Kriging model that uses a Gaussian spatial correlation function. The suggested algorithm is similar to the SCORING algorithm traditionally used in statistics.

Original languageEnglish (US)
Pages (from-to)845011-845017
Number of pages7
JournalJournal of Mechanical Design, Transactions Of the ASME
Volume131
Issue number8
DOIs
StatePublished - Aug 1 2009

Fingerprint

Maximum likelihood estimation
Linear regression
Systems analysis
Statistics

All Science Journal Classification (ASJC) codes

  • Mechanics of Materials
  • Mechanical Engineering
  • Computer Science Applications
  • Computer Graphics and Computer-Aided Design

Cite this

@article{fa8cd1189b6f473fa0e28e98f50e7e7b,
title = "Computational improvements to estimating Kriging metamodel parameters",
abstract = "The details of a method to reduce the computational burden experienced while estimating the optimal model parameters for a Kriging model are presented. A Kriging model is a type of surrogate model that can be used to create a response surface based a set of observations of a computationally expensive system design analysis. This Kriging model can then be used as a computationally efficient surrogate to the original model, providing the opportunity for the rapid exploration of the resulting tradespace. The Kriging model can provide a more complex response surface than the more traditional linear regression response surface through the introduction of a few terms to quantify the spatial correlation of the observations. Implementation details and enhancements to gradient-based methods to estimate the model parameters are presented. It concludes with a comparison of these enhancements to using maximum likelihood estimation to estimate Kriging model parameters and their potential reduction in computational burden. These enhancements include the development of the analytic gradient and Hessian for the log-likelihood equation of a Kriging model that uses a Gaussian spatial correlation function. The suggested algorithm is similar to the SCORING algorithm traditionally used in statistics.",
author = "Martin, {Jay Dean}",
year = "2009",
month = "8",
day = "1",
doi = "10.1115/1.3151807",
language = "English (US)",
volume = "131",
pages = "845011--845017",
journal = "Journal of Mechanical Design - Transactions of the ASME",
issn = "1050-0472",
publisher = "ASME",
number = "8",

}

Computational improvements to estimating Kriging metamodel parameters. / Martin, Jay Dean.

In: Journal of Mechanical Design, Transactions Of the ASME, Vol. 131, No. 8, 01.08.2009, p. 845011-845017.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Computational improvements to estimating Kriging metamodel parameters

AU - Martin, Jay Dean

PY - 2009/8/1

Y1 - 2009/8/1

N2 - The details of a method to reduce the computational burden experienced while estimating the optimal model parameters for a Kriging model are presented. A Kriging model is a type of surrogate model that can be used to create a response surface based a set of observations of a computationally expensive system design analysis. This Kriging model can then be used as a computationally efficient surrogate to the original model, providing the opportunity for the rapid exploration of the resulting tradespace. The Kriging model can provide a more complex response surface than the more traditional linear regression response surface through the introduction of a few terms to quantify the spatial correlation of the observations. Implementation details and enhancements to gradient-based methods to estimate the model parameters are presented. It concludes with a comparison of these enhancements to using maximum likelihood estimation to estimate Kriging model parameters and their potential reduction in computational burden. These enhancements include the development of the analytic gradient and Hessian for the log-likelihood equation of a Kriging model that uses a Gaussian spatial correlation function. The suggested algorithm is similar to the SCORING algorithm traditionally used in statistics.

AB - The details of a method to reduce the computational burden experienced while estimating the optimal model parameters for a Kriging model are presented. A Kriging model is a type of surrogate model that can be used to create a response surface based a set of observations of a computationally expensive system design analysis. This Kriging model can then be used as a computationally efficient surrogate to the original model, providing the opportunity for the rapid exploration of the resulting tradespace. The Kriging model can provide a more complex response surface than the more traditional linear regression response surface through the introduction of a few terms to quantify the spatial correlation of the observations. Implementation details and enhancements to gradient-based methods to estimate the model parameters are presented. It concludes with a comparison of these enhancements to using maximum likelihood estimation to estimate Kriging model parameters and their potential reduction in computational burden. These enhancements include the development of the analytic gradient and Hessian for the log-likelihood equation of a Kriging model that uses a Gaussian spatial correlation function. The suggested algorithm is similar to the SCORING algorithm traditionally used in statistics.

UR - http://www.scopus.com/inward/record.url?scp=77955220778&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77955220778&partnerID=8YFLogxK

U2 - 10.1115/1.3151807

DO - 10.1115/1.3151807

M3 - Article

VL - 131

SP - 845011

EP - 845017

JO - Journal of Mechanical Design - Transactions of the ASME

JF - Journal of Mechanical Design - Transactions of the ASME

SN - 1050-0472

IS - 8

ER -