Non-linear control variate in δ f \delta\!f particle-in-cell methods using symplectic neural networks
Victor Fournet, Martin Campos Pinto, Emmanuel Franck, Victor Michel-Dansac, [1ex] Max Planck Institute for Plasma Physics, Garching, Germany · 2026
Read Non-linear control variate in δ f \delta\!f particle-in-cell methods using symplectic neural networks by Victor Fournet, Martin Campos Pinto, Emmanuel Franck, Victor Michel-Dansac, [1ex] Max Planck Institute for Plasma Physics, Garching, Germany (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-30622.txt.
Description
Victor Fournet, Martin Campos Pinto, Emmanuel Franck and Victor Michel-Dansac present a new plasma simulation method. They simulate electrostatic plasmas with a particle-in-cell approach, which tracks representative particles instead of gridding all of phase space. Noise is the weakness: particle estimates fluctuate. The delta-f method fights noise by splitting the particle distribution into a known bulk part and a small remainder carried by particles.
The bulk density acts as a control variate, a known quantity used to cancel noise. The team evolves it with symplectic neural networks, networks designed to respect the geometry of physical motion. These networks approximate the backward flow and train on the particle trajectories themselves. A periodic variant bakes the problem's spatial repetition into the network design. Tests on the Vlasov-Poisson system, in one and three spatial dimensions plus velocities, validated the approach.
Catalog record
| CREATOR | Victor Fournet, Martin Campos Pinto, Emmanuel Franck, Victor Michel-Dansac, [1ex] Max Planck Institute for Plasma Physics, Garching, Germany |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 85 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.30622