EUVIMEDEUVIMED entdecken
Crossref

Deep learning accelerated algebraic multigrid methods for polytopal discretizations of second-order differential problems

Paola F. Antonietti, Matteo Caldana, Lorenzo Bruno Donato Gentile, Marco Verani

Mathematical Models and Methods in Applied Sciences · 2026-07 · 36 (08) · 1649-1677

DOI: 10.1142/s0218202526420029 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Mathematical Models and Methods in Applied Sciences, online erschienen am 11.04.2026. Verfasst von Paola F. Antonietti und 3 weiteren Autor:innen. Die Arbeit nennt 70 Literaturangaben. Gefördert von: European Union, MUR-PRIN/PNRR 2022.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Algebraic Multigrid (AMG) methods are state-of-the-art algebraic solvers for Partial Differential Equations. Still, their efficiency depends heavily on the choice of suitable parameters and/or ingredients. Paradigmatic examples include the so-called strong threshold parameter, which controls the algebraic coarse-grid hierarchy, as well as the smoother, i.e. the relaxation methods used on the fine grid to damp out high-frequency components of the error. In AMG, since the coarse grids are constructed algebraically (without geometric intuition), the smoother’s performance is even more critical. For the linear systems stemming from polytopal discretizations, such as Polytopal Discontinuous Galerkin (PolyDG) and Virtual Element (VEM) methods, AMG sensitivity to such choices is even more critical due to the significant variability of the underlying meshes, which results in algebraic systems with different sparsity patterns. In this paper, we focus on the linear systems of equations stemming from polytopal discretizations of second-order elliptic problems. We propose a novel deep learning approach that automatically tunes the strong threshold parameter and the smoother choice in AMG solvers, thereby maximizing AMG performance. We test various differential problems in both two- and three-dimensional settings, with heterogeneous coefficients and polygonal/polyhedral meshes, and demonstrate that the proposed approach generalizes well. In practice, we demonstrate that we can reduce AMG solver time by up to [Formula: see text] with minimal changes to existing PolyDG and VEM software libraries.

Förderung

  • European Union
  • MUR-PRIN/PNRR 2022

Zitieren

APA

Antonietti, P. F., Caldana, M., Gentile, L. B. D., & Verani, M. (2026). Deep learning accelerated algebraic multigrid methods for polytopal discretizations of second-order differential problems. Mathematical Models and Methods in Applied Sciences, 36(08), 1649-1677. https://doi.org/10.1142/s0218202526420029

Vancouver

Antonietti PF, Caldana M, Gentile LBD, Verani M. Deep learning accelerated algebraic multigrid methods for polytopal discretizations of second-order differential problems. Mathematical Models and Methods in Applied Sciences. 2026;36(08):1649-1677. doi:10.1142/s0218202526420029

BibTeX
@article{Antonietti2026,
  author = {Antonietti, Paola F. and Caldana, Matteo and Gentile, Lorenzo Bruno Donato and Verani, Marco},
  title = {Deep learning accelerated algebraic multigrid methods for polytopal discretizations of second-order differential problems},
  journal = {Mathematical Models and Methods in Applied Sciences},
  year = {2026},
  volume = {36},
  number = {08},
  pages = {1649--1677},
  doi = {10.1142/s0218202526420029}
}
Kostenlos starten →