Physics-Informed Gaussian Processes With Localized Features for Topology Optimization
Journal of Mechanical Design · 01.02.2026 · 148 (2)
DOI: 10.1115/1.4070989 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Mechanical Design, online erschienen am 12.02.2026. Verfasst von Amin Yousefpour und 3 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 30 Literaturangaben. Gefördert von: National Science Foundation, Office of Naval Research.
Abstract (Originalsprache)
We introduce a simultaneous and meshfree topology optimization (TO) framework based on physics-informed Gaussian processes (GPs). In our approach, we parameterize all design and state variables via GP priors which have a shared multi-output mean function. This mean function is represented by a customized deep neural network (DNN) whose parameters are estimated by minimizing a multi-component loss function that depends on the performance metric, design constraints, and the residuals on the state equations. Our TO approach yields well-defined material interfaces and has a built-in continuation nature that promotes global optimality. Other unique features of our approach include (1) its customized DNN which has a localized learning capacity that enables capturing intricate topologies and reducing residuals in high-gradient fields, (2) its loss function that leverages localized weights to promote solution accuracy around interfaces, and (3) its use of curriculum training to avoid local optimality. To demonstrate the power of our framework, we validate it against commercial TO package comsol on problems involving dissipated power minimization in the Stokes flow.
Förderung
- National Science Foundation
- Office of Naval Research
Zitieren
APA
Yousefpour, A., Hosseinmardi, S., Sun, X., & Bostanabad, R. (2026). Physics-Informed Gaussian Processes With Localized Features for Topology Optimization. Journal of Mechanical Design, 148(2). https://doi.org/10.1115/1.4070989
Vancouver
Yousefpour A, Hosseinmardi S, Sun X, Bostanabad R. Physics-Informed Gaussian Processes With Localized Features for Topology Optimization. Journal of Mechanical Design. 2026;148(2). doi:10.1115/1.4070989
BibTeX
@article{Yousefpour2026,
author = {Yousefpour, Amin and Hosseinmardi, Shirin and Sun, Xiangyu and Bostanabad, Ramin},
title = {Physics-Informed Gaussian Processes With Localized Features for Topology Optimization},
journal = {Journal of Mechanical Design},
year = {2026},
volume = {148},
number = {2},
doi = {10.1115/1.4070989}
}
EUVIMED