Transfer Learning in Bayesian Optimization for Aircraft Design

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Tfaily, Ali, Diouane, Youssef, Bartoli, Nathalie, Kokkolaras, Michael
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910087801470976
author Tfaily, Ali
Diouane, Youssef
Bartoli, Nathalie
Kokkolaras, Michael
author_facet Tfaily, Ali
Diouane, Youssef
Bartoli, Nathalie
Kokkolaras, Michael
contents The use of transfer learning within Bayesian optimization addresses the disadvantages of the so-called \textit{cold start} problem by using source data to aid in the optimization of a target problem. We present a method that leverages an ensemble of surrogate models using transfer learning and integrates it in a constrained Bayesian optimization framework. We identify challenges particular to aircraft design optimization related to heterogeneous design variables and constraints. We propose the use of a partial-least-squares dimension reduction algorithm to address design space heterogeneity, and a \textit{meta} data surrogate selection method to address constraint heterogeneity. Numerical benchmark problems and an aircraft conceptual design optimization problem are used to demonstrate the proposed methods. Results show significant improvement in convergence in early optimization iterations compared to standard Bayesian optimization, with improved prediction accuracy for both objective and constraint surrogate models.
format Preprint
id arxiv_https___arxiv_org_abs_2603_28999
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Transfer Learning in Bayesian Optimization for Aircraft Design
Tfaily, Ali
Diouane, Youssef
Bartoli, Nathalie
Kokkolaras, Michael
Optimization and Control
Machine Learning
The use of transfer learning within Bayesian optimization addresses the disadvantages of the so-called \textit{cold start} problem by using source data to aid in the optimization of a target problem. We present a method that leverages an ensemble of surrogate models using transfer learning and integrates it in a constrained Bayesian optimization framework. We identify challenges particular to aircraft design optimization related to heterogeneous design variables and constraints. We propose the use of a partial-least-squares dimension reduction algorithm to address design space heterogeneity, and a \textit{meta} data surrogate selection method to address constraint heterogeneity. Numerical benchmark problems and an aircraft conceptual design optimization problem are used to demonstrate the proposed methods. Results show significant improvement in convergence in early optimization iterations compared to standard Bayesian optimization, with improved prediction accuracy for both objective and constraint surrogate models.
title Transfer Learning in Bayesian Optimization for Aircraft Design
topic Optimization and Control
Machine Learning
url https://arxiv.org/abs/2603.28999