Generalized integrated brownian fields for simulation metamodeling

Peter Salemi, Jeremy Staum, Barry L. Nelson

Research output: Chapter in Book/Report/Conference proceedingConference contribution

10 Scopus citations

Abstract

We use Gaussian random fields (GRFs) that we call generalized integrated Brownian fields (GIBFs), whose covariance functions have been studied in the context of reproducing kernels, for Gaussian process modeling. We introduce GIBFs into the fields of deterministic and stochastic simulation metamodeling, and give a probabilistic representation of GIBFs that is not given in the literature on reproducing kernels. These GIBFs have differentiability that can be controlled in each coordinate, and are built from GRFs which have the Markov property. Furthermore, we introduce a new parameterization of GIBFs which allows them to be used in higher-dimensional metamodeling problems. We also show how to implement stochastic kriging with GIBFs, covering trend modeling and fitting. Lastly, we use tractable examples to demonstrate superior prediction ability as compared to the GRF corresponding to the Gaussian covariance function.

Original languageEnglish (US)
Title of host publicationProceedings of the 2013 Winter Simulation Conference - Simulation
Subtitle of host publicationMaking Decisions in a Complex World, WSC 2013
Pages543-554
Number of pages12
DOIs
StatePublished - 2013
Event2013 43rd Winter Simulation Conference - Simulation: Making Decisions in a Complex World, WSC 2013 - Washington, DC, United States
Duration: Dec 8 2013Dec 11 2013

Publication series

NameProceedings of the 2013 Winter Simulation Conference - Simulation: Making Decisions in a Complex World, WSC 2013

Other

Other2013 43rd Winter Simulation Conference - Simulation: Making Decisions in a Complex World, WSC 2013
CountryUnited States
CityWashington, DC
Period12/8/1312/11/13

ASJC Scopus subject areas

  • Modeling and Simulation

Fingerprint Dive into the research topics of 'Generalized integrated brownian fields for simulation metamodeling'. Together they form a unique fingerprint.

Cite this