Leveraging Operator Learning to Accelerate Convergence of the Preconditioned Conjugate Gradient Method

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Kopaničáková, Alena, Lee, Youngkyu, Karniadakis, George Em
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866913969365581824
author Kopaničáková, Alena
Lee, Youngkyu
Karniadakis, George Em
author_facet Kopaničáková, Alena
Lee, Youngkyu
Karniadakis, George Em
contents We propose a new deflation strategy to accelerate the convergence of the preconditioned conjugate gradient(PCG) method for solving parametric large-scale linear systems of equations. Unlike traditional deflation techniques that rely on eigenvector approximations or recycled Krylov subspaces, we generate the deflation subspaces using operator learning, specifically the Deep Operator Network~(DeepONet). To this aim, we introduce two complementary approaches for assembling the deflation operators. The first approach approximates near-null space vectors of the discrete PDE operator using the basis functions learned by the DeepONet. The second approach directly leverages solutions predicted by the DeepONet. To further enhance convergence, we also propose several strategies for prescribing the sparsity pattern of the deflation operator. A comprehensive set of numerical experiments encompassing steady-state, time-dependent, scalar, and vector-valued problems posed on both structured and unstructured geometries is presented and demonstrates the effectiveness of the proposed DeepONet-based deflated PCG method, as well as its generalization across a wide range of model parameters and problem resolutions.
format Preprint
id arxiv_https___arxiv_org_abs_2508_00101
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Leveraging Operator Learning to Accelerate Convergence of the Preconditioned Conjugate Gradient Method
Kopaničáková, Alena
Lee, Youngkyu
Karniadakis, George Em
Numerical Analysis
Machine Learning
Optimization and Control
65M55, 68T05, 49K20
We propose a new deflation strategy to accelerate the convergence of the preconditioned conjugate gradient(PCG) method for solving parametric large-scale linear systems of equations. Unlike traditional deflation techniques that rely on eigenvector approximations or recycled Krylov subspaces, we generate the deflation subspaces using operator learning, specifically the Deep Operator Network~(DeepONet). To this aim, we introduce two complementary approaches for assembling the deflation operators. The first approach approximates near-null space vectors of the discrete PDE operator using the basis functions learned by the DeepONet. The second approach directly leverages solutions predicted by the DeepONet. To further enhance convergence, we also propose several strategies for prescribing the sparsity pattern of the deflation operator. A comprehensive set of numerical experiments encompassing steady-state, time-dependent, scalar, and vector-valued problems posed on both structured and unstructured geometries is presented and demonstrates the effectiveness of the proposed DeepONet-based deflated PCG method, as well as its generalization across a wide range of model parameters and problem resolutions.
title Leveraging Operator Learning to Accelerate Convergence of the Preconditioned Conjugate Gradient Method
topic Numerical Analysis
Machine Learning
Optimization and Control
65M55, 68T05, 49K20
url https://arxiv.org/abs/2508.00101