Bayesian Optimization with Setup Switching Cost

Stefan Pricopie; Richard Allmendinger; Manuel López-Ibáñez; Clyde Fare; Matt Benatan; Joshua Knowles

doi:10.1145/3638530.3664101

Bayesian Optimization with Setup Switching Cost

Stefan Pricopie^*, Richard Allmendinger, Manuel López-Ibáñez, Clyde Fare, Matt Benatan, Joshua Knowles

^*Corresponding author for this work

Research output: Contribution to conference › Abstract › peer-review

Abstract

Bayesian Optimization (BO) in its classical form is cost-unaware. However, many real-world problems are resource-constrained and hence incur a cost whenever such resources are needed, such as when a new setup is used. We are then looking at adapted cost-aware solution methods that are improving the performance of BO over cost-constrained problems. We find that parameter-free algorithms can yield comparable results to fine-tuned algorithms used in constrained optimization.

Original language	English
Pages	103-104
Number of pages	2
DOIs	https://doi.org/10.1145/3638530.3664101
Publication status	Published - 14 Jul 2024
Event	2024 Genetic and Evolutionary Computation Conference Companion, GECCO 2024 Companion - Melbourne, Australia Duration: 14 Jul 2024 → 18 Jul 2024

Conference

Conference	2024 Genetic and Evolutionary Computation Conference Companion, GECCO 2024 Companion
Country/Territory	Australia
City	Melbourne
Period	14/07/24 → 18/07/24

Keywords

bayesian optimization
expensive optimization
switching cost

Access to Document

10.1145/3638530.3664101

Cite this

@conference{5262b4a8a61e44e8a4b1ddaa22d22827,

title = "Bayesian Optimization with Setup Switching Cost",

abstract = "Bayesian Optimization (BO) in its classical form is cost-unaware. However, many real-world problems are resource-constrained and hence incur a cost whenever such resources are needed, such as when a new setup is used. We are then looking at adapted cost-aware solution methods that are improving the performance of BO over cost-constrained problems. We find that parameter-free algorithms can yield comparable results to fine-tuned algorithms used in constrained optimization.",

keywords = "bayesian optimization, expensive optimization, switching cost",

author = "Stefan Pricopie and Richard Allmendinger and Manuel L{\'o}pez-Ib{\'a}{\~n}ez and Clyde Fare and Matt Benatan and Joshua Knowles",

note = "Publisher Copyright: {\textcopyright} 2024 Copyright held by the owner/author(s).; 2024 Genetic and Evolutionary Computation Conference Companion, GECCO 2024 Companion ; Conference date: 14-07-2024 Through 18-07-2024",

year = "2024",

month = jul,

day = "14",

doi = "10.1145/3638530.3664101",

language = "English",

pages = "103--104",

}

TY - CONF

T1 - Bayesian Optimization with Setup Switching Cost

AU - Pricopie, Stefan

AU - Allmendinger, Richard

AU - López-Ibáñez, Manuel

AU - Fare, Clyde

AU - Benatan, Matt

AU - Knowles, Joshua

PY - 2024/7/14

Y1 - 2024/7/14

N2 - Bayesian Optimization (BO) in its classical form is cost-unaware. However, many real-world problems are resource-constrained and hence incur a cost whenever such resources are needed, such as when a new setup is used. We are then looking at adapted cost-aware solution methods that are improving the performance of BO over cost-constrained problems. We find that parameter-free algorithms can yield comparable results to fine-tuned algorithms used in constrained optimization.

AB - Bayesian Optimization (BO) in its classical form is cost-unaware. However, many real-world problems are resource-constrained and hence incur a cost whenever such resources are needed, such as when a new setup is used. We are then looking at adapted cost-aware solution methods that are improving the performance of BO over cost-constrained problems. We find that parameter-free algorithms can yield comparable results to fine-tuned algorithms used in constrained optimization.

KW - bayesian optimization

KW - expensive optimization

KW - switching cost

UR - http://www.scopus.com/inward/record.url?scp=85201964823&partnerID=8YFLogxK

U2 - 10.1145/3638530.3664101

DO - 10.1145/3638530.3664101

M3 - Abstract

AN - SCOPUS:85201964823

SP - 103

EP - 104

T2 - 2024 Genetic and Evolutionary Computation Conference Companion, GECCO 2024 Companion

Y2 - 14 July 2024 through 18 July 2024

ER -

Bayesian Optimization with Setup Switching Cost

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this