# Adaptive hybrid coupling with operator inference, the overlapping Schwarz alternating method and reinforcement learning

DevFeed: [Adaptive hybrid coupling with operator inference, the overlapping Schwarz alternating method and reinforcement learning](<https://devfeed.tech/articles/adaptive-hybrid-coupling-with-operator-inference-the-overlapping-schwarz-alternating-method-and-reinforcement-learning-108535.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.17837>)

Author: Trishit Mondal, Irina Tezaur, Anthony Gruber

Published: 2026-10-30T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Reinforcement learning](<https://devfeed.tech/topics/reinforcement-learning.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>)

Tags: [3d](<https://devfeed.tech/tags/3d.md>), [adaptive](<https://devfeed.tech/tags/adaptive.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [coupling](<https://devfeed.tech/tags/coupling.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [domain-decomposition](<https://devfeed.tech/tags/domain-decomposition.md>), [elastic](<https://devfeed.tech/tags/elastic.md>), [hybrid](<https://devfeed.tech/tags/hybrid.md>), [inference](<https://devfeed.tech/tags/inference.md>), [math-mp](<https://devfeed.tech/tags/math-mp.md>), [math-ph](<https://devfeed.tech/tags/math-ph.md>), [reduced-order-models](<https://devfeed.tech/tags/reduced-order-models.md>), [reinforcement-learning](<https://devfeed.tech/tags/reinforcement-learning.md>), [rl](<https://devfeed.tech/tags/rl.md>), [sam](<https://devfeed.tech/tags/sam.md>)

## AI overview

Deep Q-networks can adapt model fidelity during overlapping Schwarz alternating method (O-SAM) simulations, assigning full order models (FOMs) or pre-trained Operator Inference (OpInf) reduced order models (ROMs) to subdomains as features move. Trained offline with a reward balancing accuracy, cost, and switching frequency, policies are deployed on unseen problems without a reference FOM solution. In a 1D moving-front advection-diffusion case, dynamic allocation outperformed static assignments; allowing the agent to adapt the decomposition brought no further benefit. In 3D elastic wave propagation, policies for two- and three-subdomain decompositions assigned FOMs where the wave was present and ROMs elsewhere. The abstract presents this as a demonstrated potential for predictive online fidelity adaptation in Schwarz-based hybrid simulations.

## Source excerpt

arXiv:2609.17837v2 Announce Type: replace Abstract: Hybrid domain decomposition methods provide a flexible framework for coupling full order models (FOMs) and reduced order models (ROMs), but typically assume the model assigned to each subdomain is fixed throughout a simulation. This is limiting for transient problems in which localized features propagate through the domain and the regions requiring high-fidelity resolution change over time. We introduce a reinforcement learning (RL)-based approach for online adaptation of FOM-ROM models coupled via the overlapping Schwarz alternating method (O-SAM), an iterative domain decomposition method that solves subdomain-local problems while exchanging solution information through transmission boundary conditions on overlapping interfaces. Deep Q-networks (DQNs) are trained offline to select among subdomain-local FOMs and pre-trained Operator Inference (OpInf) ROMs using a reward balancing accuracy, cost, and model-switching frequency. Once trained, the policies are deployed predictively on problem instances not seen during training, without requiring a reference FOM solution. We demonstrate the approach on two examples: a 1D advection-diffusion problem with a moving front, and a 3D linear elastic wave propagation problem implemented in the Norma.jl solid mechanics code. For the advection-diffusion benchmark, the learned policy dynamically allocates high-fidelity resolution as the front propagates and outperforms static FOM/ROM assignments; letting the agent also adapt the domain decomposition provides no further benefit. For the elastic wave benchmark, learned policies for two and three subdomain decompositions track the propagating wave by assigning FOMs to subdomains containing the wave and ROMs elsewhere, as expected. Our results demonstrate the potential of RL to enable predictive online adaptation of model fidelity within Schwarz-based hybrid simulations.