Selective boundary condition reduction via learned error gating

DOI10.48550/arXiv.2609.08461MaRDI QIDQ7288239arXivFDO

Authors Daniel Fernández, Dominik Penk, Dominik Riedelbauch

Publication date 8 September 2026

Published in arXiv

Tldr This paper presents a framework for learning when to use simplified boundary conditions in parametric PDEs, reducing computational cost while maintaining accuracy: neural networks estimate domain and boundary errors of paired solutions, replacing the more expensive condition with the simpler one if both predicted errors meet prescribed tolerances. We focus on singular limits and evaluate the method on a galvanic corrosion problem and other nonlinear stationary and evolution problems.

Full work available at URL https://doi.org/10.48550/arXiv.2609.08461

Mathematics Subject Classification ID

Singular perturbations in context of PDEs (35B25) Boundary value problems for second-order elliptic equations (35J25) Artificial neural networks and deep learning (68T07)

arXiv classification

cs.LGmath.NA

Cites work

• A New Certified Hierarchical and Adaptive RB-ML-ROM Surrogate Model for Parametrized PDEs
• Accurate and robust predictions for model order reduction via an adaptive, hybrid FOM/ROM approach
• Distribution-free predictive inference for regression
• Electrode modelling: the effect of contact impedance
• Machine-learning error models for approximate solutions to parameterized systems of nonlinear equations
• Nitsche's method for general boundary conditions
• Nonlinear Reconstruction for Operator Learning of PDEs with Discontinuities
• On optimum recognition error and reject tradeoff
• Robin approximation of Dirichlet boundary value problems
• Semi-analytic PINN methods for singularly perturbed boundary value problems

This page was built for publication: Selective boundary condition reduction via learned error gating

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7288239)