Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A collision-oriented interacting particle system for Landau-type equations and the molecular chaos

Kai Du, Lei Li

TL;DR

This work introduces a collision-oriented particle system that mimics grazing-collision dynamics to approximate Landau-type equations with drift kernel $K$ and diffusion kernel $A$. The system employs pairwise diffusion within randomly formed $N/2$ groups at each time step, yielding an intrinsic connection to the grazing limit and a computational cost of $O(N)$ per step. The authors establish a quantitative mean-field limit using relative entropy, proving propagation of chaos and explicit rates that depend on the time step $ riangle t$ and the particle number $N$, under regularity assumptions on $K$ and $A$. Central to the analysis are Li–Yau type gradient estimates for the mean-field equation and integral estimates for derivatives of the logarithmic densities of the particle system, ensuring the diffusion term is well-approximated by mean-field interactions. The results justify the microscopic modeling choice, show enhanced chaos with higher collision frequency, and point to practical numerical advantages for simulating Landau-type dynamics, with future directions including numerical implementation and extensions to true Landau settings.

Abstract

We propose a collision-oriented particle system to approximate a class of Landau-type equations. This particle system is formally derived from a particle system with random collisions in the grazing regime, and happens to be a special random batch system with random interaction in the diffusion coefficient. The difference from usual random batch systems with random interaction in the drift is that the batch size has to be $p=2$. We then analyze the convergence rate of the proposed particle system to the Landau-type equations using the tool of relative entropy, assuming that the interaction kernels are regular enough. A key aspect of our approach is the gradient estimates of logarithmic densities, applied to both the Landau-type equations and the particle systems. Compared to existing particle systems for the approximation of Landau-type equations, our proposed system not only offers a more intrinsic reflection of the underlying physics but also reduces the computational cost to $O(N)$ per time step when implemented numerically.

A collision-oriented interacting particle system for Landau-type equations and the molecular chaos

TL;DR

This work introduces a collision-oriented particle system that mimics grazing-collision dynamics to approximate Landau-type equations with drift kernel and diffusion kernel . The system employs pairwise diffusion within randomly formed groups at each time step, yielding an intrinsic connection to the grazing limit and a computational cost of per step. The authors establish a quantitative mean-field limit using relative entropy, proving propagation of chaos and explicit rates that depend on the time step and the particle number , under regularity assumptions on and . Central to the analysis are Li–Yau type gradient estimates for the mean-field equation and integral estimates for derivatives of the logarithmic densities of the particle system, ensuring the diffusion term is well-approximated by mean-field interactions. The results justify the microscopic modeling choice, show enhanced chaos with higher collision frequency, and point to practical numerical advantages for simulating Landau-type dynamics, with future directions including numerical implementation and extensions to true Landau settings.

Abstract

We propose a collision-oriented particle system to approximate a class of Landau-type equations. This particle system is formally derived from a particle system with random collisions in the grazing regime, and happens to be a special random batch system with random interaction in the diffusion coefficient. The difference from usual random batch systems with random interaction in the drift is that the batch size has to be . We then analyze the convergence rate of the proposed particle system to the Landau-type equations using the tool of relative entropy, assuming that the interaction kernels are regular enough. A key aspect of our approach is the gradient estimates of logarithmic densities, applied to both the Landau-type equations and the particle systems. Compared to existing particle systems for the approximation of Landau-type equations, our proposed system not only offers a more intrinsic reflection of the underlying physics but also reduces the computational cost to per time step when implemented numerically.
Paper Structure (17 sections, 13 theorems, 200 equations, 4 algorithms)

This paper contains 17 sections, 13 theorems, 200 equations, 4 algorithms.

Table of Contents

  1. Introduction
  2. A formal derivation for the collision-oriented particle system
  3. A particle system in the velocity space for Kac equation
  4. The grazing regime and the particle systems
  5. Approximation error and propagation of chaos
  6. Assumptions, main results and discussions
  7. Proof of the main results
  8. The solution to the mean field equation
  9. Estimates of the density for the particle system
  10. Main estimates for derivatives of logarithmic density
  11. The proofs for the estimates of the particle density
  12. Estimates for the local error terms
  13. Closing the loose ends
  14. The concentration estimate
  15. Mean field approximation of the diffusion coefficient
  16. ...and 2 more sections

Key Result

Lemma 3.1

For symmetric measures, if $1\le k\le \ell\le N$, one has

Theorems & Definitions (30)

  • Remark 2.1
  • Remark 2.2
  • Remark 2.3
  • Definition 3.1
  • Lemma 3.1
  • Theorem 3.1
  • Definition 3.2
  • Lemma 3.2
  • Lemma 3.3
  • Remark 3.1
  • ...and 20 more