Multiscale Mixed-Dimensional Simulation via Domain Decomposition and Non-Intrusive Neural Model Order Reduction

DOI10.48550/arXiv.2607.15171MaRDI QIDQ7284080arXivFDO

Authors Nunzio Dimola, Piermario Vitullo, Paolo Zunino

Publication date 16 July 2026

Published in arXiv

Tldr This paper introduces a Domain Decomposition Reduced Order Model (DD-ROM) for multiscale simulations via non-intrusive neural model order reduction, enabling efficient and accurate solutions of mixed-dimensional problems with complex geometries and parameters. The key contribution is an iterative solver that replaces the fine-scale local operators by their neural surrogates learned from affordable local data, ensuring stability and good scalability.

Full work available at URL https://doi.org/10.48550/arXiv.2607.15171

Mathematics Subject Classification ID

Multigrid methods; domain decomposition for initial value and initial-boundary value problems involving PDEs (65M55) Finite element, Rayleigh-Ritz and Galerkin methods for boundary value problems involving PDEs (65N30) Multigrid methods; domain decomposition for boundary value problems involving PDEs (65N55) Artificial neural networks and deep learning (68T07)

arXiv classification

math.NA

Cites work

• A Mathematical Guide to Operator Learning
• A non-overlapping optimization-based domain decomposition approach to component-based model reduction of incompressible flows
• A survey of projection-based model reduction methods for parametric dynamical systems
• A Theoretical Analysis of Deep Neural Networks and Parametric PDEs
• Analysis and approximation of mixed-dimensional PDEs on 3D-1D domains coupled with Lagrange multipliers
• Asymptotic analysis for periodic structures
• Best Robin parameters for optimized Schwarz methods at cross points
• Combining machine learning and domain decomposition methods for the solution of partial differential equations—A review
• Combining physics-based and data-driven models: advancing the frontiers of research with scientific machine learning
• Derivation and analysis of coupled PDEs on manifolds with high dimensionality gap arising from topological model reduction
• Domain decomposition for data-driven reduced modeling of large-scale systems
• Domain decomposition methods
• Hybrid models for simulating blood flow in microvascular networks
• Introduction to the Combination of Reduced Order Models and Domain Decomposition: State of the Art and Perspectives
• Localized model order reduction and domain decomposition methods for coupled heterogeneous systems
• Localized orthogonal decomposition techniques for boundary value problems
• Localized orthogonal decomposition techniques for boundary value problems
• Mesh-informed neural networks for operator learning in finite element spaces
• Modeling flow in porous media with fractures; discrete fracture models with matrix-fracture exchange
• Multidimensional coupling: a variationally consistent approach to fiber-reinforced materials
• Non-intrusive reduced order modeling of nonlinear problems using neural networks
• Nonlinear model order reduction for problems with microstructure using mesh informed neural networks
• Numerical homogenization of spatial network models
• On the limited memory BFGS method for large scale optimization
• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• Reduced basis methods for partial differential equations. An introduction
• Reduced Lagrange multiplier approach for non-matching coupling of mixed-dimensional domains
• The heterogeneous multiscale methods

This page was built for publication: Multiscale Mixed-Dimensional Simulation via Domain Decomposition and Non-Intrusive Neural Model Order Reduction

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7284080)