Gaussian Processes for Simulation-Based Optimization and Robust Design

Alex Durkin; Marcos Millan-Agorio; Miao Guo

doi:10.1016/B978-0-323-85159-6.50207-4

Gaussian Processes for Simulation-Based Optimization and Robust Design

Alex Durkin
, Marcos Millan-Agorio
, Miao Guo

Research output: Chapter in Book/Report/Conference proceeding › Chapter › peer-review

3 Scopus citations

Abstract

Gaussian Processes present a versatile surrogate modeling toolbox to address simulation-based optimization and uncertainties arising from non-converged simulations. In this work we present a black-box optimization methodology framework in which Gaussian Process Regression is used to model complex underlying process performance models and Gaussian Process Classification is used to model feasibility constraints based on converged and non-converged simulations. Additionally, we present a conservativeness parameter to enable tuning of the feasible region based on the trade-off between process performance and the risk of infeasibility due to non-converged simulations.

Original language	English
Title of host publication	Computer Aided Chemical Engineering
Publisher	Elsevier B.V.
Pages	1243-1248
Number of pages	6
DOIs	https://doi.org/10.1016/B978-0-323-85159-6.50207-4
State	Published - Jan 2022
Externally published	Yes

Publication series

Name	Computer Aided Chemical Engineering
Volume	49
ISSN (Print)	1570-7946

Bibliographical note

Publisher Copyright:
© 2022 Elsevier B.V.

Keywords

Gaussian Processes
Optimization
Surrogate Modeling

ASJC Scopus subject areas

General Chemical Engineering
Computer Science Applications

Access to Document

10.1016/B978-0-323-85159-6.50207-4

Cite this

@inbook{9cbd12ca9ca64b41b97bcae02553a10a,

title = "Gaussian Processes for Simulation-Based Optimization and Robust Design",

abstract = "Gaussian Processes present a versatile surrogate modeling toolbox to address simulation-based optimization and uncertainties arising from non-converged simulations. In this work we present a black-box optimization methodology framework in which Gaussian Process Regression is used to model complex underlying process performance models and Gaussian Process Classification is used to model feasibility constraints based on converged and non-converged simulations. Additionally, we present a conservativeness parameter to enable tuning of the feasible region based on the trade-off between process performance and the risk of infeasibility due to non-converged simulations.",

keywords = "Gaussian Processes, Optimization, Surrogate Modeling",

author = "Alex Durkin and Marcos Millan-Agorio and Miao Guo",

note = "Publisher Copyright: {\textcopyright} 2022 Elsevier B.V.",

year = "2022",

month = jan,

doi = "10.1016/B978-0-323-85159-6.50207-4",

language = "English",

series = "Computer Aided Chemical Engineering",

publisher = "Elsevier B.V.",

pages = "1243--1248",

booktitle = "Computer Aided Chemical Engineering",

}

TY - CHAP

T1 - Gaussian Processes for Simulation-Based Optimization and Robust Design

AU - Durkin, Alex

AU - Millan-Agorio, Marcos

AU - Guo, Miao

PY - 2022/1

Y1 - 2022/1

N2 - Gaussian Processes present a versatile surrogate modeling toolbox to address simulation-based optimization and uncertainties arising from non-converged simulations. In this work we present a black-box optimization methodology framework in which Gaussian Process Regression is used to model complex underlying process performance models and Gaussian Process Classification is used to model feasibility constraints based on converged and non-converged simulations. Additionally, we present a conservativeness parameter to enable tuning of the feasible region based on the trade-off between process performance and the risk of infeasibility due to non-converged simulations.

AB - Gaussian Processes present a versatile surrogate modeling toolbox to address simulation-based optimization and uncertainties arising from non-converged simulations. In this work we present a black-box optimization methodology framework in which Gaussian Process Regression is used to model complex underlying process performance models and Gaussian Process Classification is used to model feasibility constraints based on converged and non-converged simulations. Additionally, we present a conservativeness parameter to enable tuning of the feasible region based on the trade-off between process performance and the risk of infeasibility due to non-converged simulations.

KW - Gaussian Processes

KW - Optimization

KW - Surrogate Modeling

UR - https://www.scopus.com/pages/publications/85136241724

U2 - 10.1016/B978-0-323-85159-6.50207-4

DO - 10.1016/B978-0-323-85159-6.50207-4

M3 - Chapter

AN - SCOPUS:85136241724

T3 - Computer Aided Chemical Engineering

SP - 1243

EP - 1248

BT - Computer Aided Chemical Engineering

PB - Elsevier B.V.

ER -

Gaussian Processes for Simulation-Based Optimization and Robust Design

Abstract

Publication series

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this