A Frobenius-Optimal Projection for Enforcing Linear Conservation in Learned Dynamical Models

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Mango, John M., Katende, Ronald
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912789869625344
author Mango, John M.
Katende, Ronald
author_facet Mango, John M.
Katende, Ronald
contents We consider the problem of restoring linear conservation laws in data-driven linear dynamical models. Given a learned operator $\widehat{A}$ and a full-rank constraint matrix $C$ encoding one or more invariants, we show that the matrix closest to $\widehat{A}$ in the Frobenius norm and satisfying $C^\top A = 0$ is the orthogonal projection $A^\star = \widehat{A} - C(C^\top C)^{-1}C^\top \widehat{A}$. This correction is uniquely defined, low rank and fully determined by the violation $C^\top \widehat{A}$. In the single-invariant case it reduces to a rank-one update. We prove that $A^\star$ enforces exact conservation while minimally perturbing the dynamics, and we verify these properties numerically on a Markov-type example. The projection provides an elementary and general mechanism for embedding exact invariants into any learned linear model.
format Preprint
id arxiv_https___arxiv_org_abs_2512_22084
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle A Frobenius-Optimal Projection for Enforcing Linear Conservation in Learned Dynamical Models
Mango, John M.
Katende, Ronald
Dynamical Systems
Machine Learning
Numerical Analysis
We consider the problem of restoring linear conservation laws in data-driven linear dynamical models. Given a learned operator $\widehat{A}$ and a full-rank constraint matrix $C$ encoding one or more invariants, we show that the matrix closest to $\widehat{A}$ in the Frobenius norm and satisfying $C^\top A = 0$ is the orthogonal projection $A^\star = \widehat{A} - C(C^\top C)^{-1}C^\top \widehat{A}$. This correction is uniquely defined, low rank and fully determined by the violation $C^\top \widehat{A}$. In the single-invariant case it reduces to a rank-one update. We prove that $A^\star$ enforces exact conservation while minimally perturbing the dynamics, and we verify these properties numerically on a Markov-type example. The projection provides an elementary and general mechanism for embedding exact invariants into any learned linear model.
title A Frobenius-Optimal Projection for Enforcing Linear Conservation in Learned Dynamical Models
topic Dynamical Systems
Machine Learning
Numerical Analysis
url https://arxiv.org/abs/2512.22084