Neural Network-Assisted Refinement of Traditional Schemes for One-Dimensional Scalar Conservation Laws

DOI10.1007/S10444-026-10357-WMaRDI QIDQ7360689arXivFDO

Authors Ferenc Izsák, Imre Fekete, Vendel P. Kupás

Publication date 26 September 2026

Published in arXiv

Tldr We develop neural networks that refine traditional numerical schemes for solving one-dimensional scalar conservation laws by learning improved flux functions and slope-limiter terms while maintaining low model complexity and enabling sequential training for multi-step time-stepping.

Full work available at URL https://doi.org/10.1007/S10444-026-10357-W

arXiv classification

cs.LGmath.NA

Cites work

• First order quasilinear equations with boundary conditions
• High resolution schemes for hyperbolic conservation laws
• Learning the Dynamics for Unknown Hyperbolic Conservation Laws Using Deep Neural Networks
• Machine learning-based WENO5 scheme
• wPINNs: Weak Physics Informed Neural Networks for Approximating Entropy Solutions of Hyperbolic Conservation Laws

This page was built for publication: Neural Network-Assisted Refinement of Traditional Schemes for One-Dimensional Scalar Conservation Laws

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7360689)