Penalization method to convert Bayesian optimization methods into batch multi-objective Bayesian optimization methods

Adelle Holder, Henry DeBruin, Jesse M Sestito · 2026

Read Penalization method to convert Bayesian optimization methods into batch multi-objective Bayesian optimization methods by Adelle Holder, Henry DeBruin, Jesse M Sestito (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0354346.txt.

Description

Adelle Holder, Henry DeBruin, and Jesse Sestito tackle a bottleneck in computer-aided design in this 2026 PLOS ONE methods paper. Bayesian optimisation searches for good designs by testing one candidate at a time, which wastes time when many tests could run in parallel. Existing parallel versions are usually built from scratch, ignoring a large library of proven one-at-a-time methods.

Their fix wraps any compatible sequential method in a penalty scheme that steers each batch of tests apart, so the batch covers fresh ground instead of repeating itself. Two converted methods matched established parallel tools on solution quality and beat them on real time. Designers gain a shortcut: upgrade familiar methods rather than rebuild them.

Catalog record

CREATORAdelle Holder, Henry DeBruin, Jesse M Sestito
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE41638
LICENSECC_BY
RIGHTSCC BY
SOURCEPLOS

Source: https://doi.org/10.1371/journal.pone.0354346

Browse the full library · About the Tower of Babel