An adaptive Gaussian splitting scheme for linear Fokker--Planck equations

We develop an adaptive numerical method for linear Fokker--Planck equations on $\mathbb R^d$, based on approximating the evolving probability density by linear combinations of Gaussian functions (LCGs). The method applies to polynomial-type drift fields, without requiring a gradient-flow structure or a stationary distribution. It relies on an operator splitting that separates diffusion and affine drift, for which Gaussian functions can be propagated exactly, from higher-order transport, whose evolution is approximated adaptively by LCGs. Both the number of Gaussian components and their means and covariances are adjusted to meet the prescribed accuracy tolerances. The adaptivity is driven by computable error indicators involving polynomial moments of Gaussians that can be evaluated analytically. We provide a rigorous error analysis that relates the approximation error to the indicators driving adaptivity and illustrate the performance of the scheme in several test cases.

Authors

Publication Details

Journal
arXiv (Cornell University)
Published
2026-11-16
DOI
https://doi.org/10.48550/arxiv.2611.20743
Primary Topic
Numerical Solvers for Ordinary Differential Equations
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

An adaptive Gaussian splitting scheme for linear Fokker--Planck equations

Markus Bachmayr, Damiano Lombardi, Tianyu Jin, Ahmad Zorkot et al.
arXiv (Cornell University)
Numerical Solvers for Ordinary Differential Equations
preprint

An adaptive Gaussian splitting scheme for linear Fokker--Planck equations

Markus Bachmayr, Damiano Lombardi, Tianyu Jin, Ahmad Zorkot, Olga Mula
preprint en

Abstract

We develop an adaptive numerical method for linear Fokker--Planck equations on $\mathbb R^d$, based on approximating the evolving probability density by linear combinations of Gaussian functions (LCGs). The method applies to polynomial-type drift fields, without requiring a gradient-flow structure or a stationary distribution. It relies on an operator splitting that separates diffusion and affine drift, for which Gaussian functions can be propagated exactly, from higher-order transport, whose evolution is approximated adaptively by LCGs. Both the number of Gaussian components and their means and covariances are adjusted to meet the prescribed accuracy tolerances. The adaptivity is driven by computable error indicators involving polynomial moments of Gaussians that can be evaluated analytically. We provide a rigorous error analysis that relates the approximation error to the indicators driving adaptivity and illustrate the performance of the scheme in several test cases.

arXiv (Cornell University)
Numerical Solvers for Ordinary Differential Equations
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon