A note on asymptotics of linear dissipative kinetic equations in bounded domains

Basic Information

• Paper ID: 2309.15758
• Title: A note on asymptotics of linear dissipative kinetic equations in bounded domains
• Author: Yuzhe Zhu
• Classification: math.AP (Analysis of PDEs)
• Publication Date: October 10, 2024 (v2)
• Paper Link: https://arxiv.org/abs/2309.15758

Abstract

This paper establishes $L^2$ exponential decay properties for linear dissipative kinetic equations (including time relaxation and Fokker-Planck models) on bounded spatial domains with general boundary conditions that may not conserve mass. Diffusive asymptotics are derived under general Maxwell boundary conditions. The proof methodology is based on energy estimates, combining ideas from $L^2$-hypocoercivity and relative entropy methods.

Research Background and Motivation

Problem Description

This paper studies linear dissipative kinetic equations with parameter $\varepsilon \in (0,1]$: $\varepsilon\partial_t f_\varepsilon + v \cdot \nabla_x f_\varepsilon - \nabla_x \phi \cdot \nabla_v f_\varepsilon = \varepsilon^{-1} L_i f_\varepsilon$

where $L_1$ is the time relaxation operator, $L_2$ is the Fokker-Planck operator, equipped with general boundary conditions $f_\varepsilon = \alpha D f_\varepsilon + \beta R f_\varepsilon$.

Research Motivation

1. Theoretical necessity: Existing $L^2$-hypocoercivity theory primarily addresses periodic domains or whole spaces, with limited treatment of bounded domains
2. Generality of boundary conditions: Most existing work assumes mass-conserving boundary conditions; this paper considers more general non-conservative cases
3. Diffusive limits: Understanding the macroscopic diffusive behavior of kinetic equations as $\varepsilon \to 0$
4. Applied value: Such equations describe particle system evolution in external force fields, with important applications in neutron transport and plasma physics

Core Contributions

1. Established unified $L^2$ exponential decay estimates: For both time relaxation and Fokker-Planck operators, obtained $\varepsilon$-uniform exponential decay rates under general boundary conditions
2. Handled non-conservative boundary conditions: When $\alpha + \beta < 1$, the system does not conserve mass, obtaining complete decay of solutions to zero
3. Proved diffusive asymptotics: Under Maxwell boundary conditions, demonstrated that as $\varepsilon \to 0$, solutions converge to solutions of parabolic equations with Neumann boundary conditions
4. Provided quantitative convergence rates: Gave explicit $O(\sqrt{\varepsilon})$ convergence rates for different initial value regularities

Detailed Methodology

Problem Formulation

Study kinetic equations on bounded domain $\Omega \subset \mathbb{R}^d$:

• Input: Initial value $f_{in} \in L^2(\Omega \times \mathbb{R}^d, dm)$, boundary parameters $\alpha, \beta$
• Output: Long-time asymptotic behavior and diffusive limits of solutions
• Constraints: $\alpha, \beta: \partial\Omega \to [0,1]$, $\alpha + \beta \leq 1$

Core Technical Framework

1. Boundary Term Treatment Technique

Lemmas 2.1 and 2.2 provide key identities for handling general boundary conditions: $\left(f^2, n_x \cdot v\right)_\partial = \left((1-\beta^2)(f-Df)^2, (n_x \cdot v)_+\right)_\partial + \left((1-(\alpha+\beta)^2)(Df)^2, (n_x \cdot v)_+\right)_\partial$

The key aspect of this identity is decomposing boundary terms into non-negative parts, providing dissipative structure for energy estimates.

2. Modified Entropy Method

Construct modified entropy function: $E_\varepsilon := \|f_\varepsilon - M_c\|^2 + \kappa\varepsilon(v \cdot \nabla_x u, f_\varepsilon^\perp)$

where $u$ is the solution to an elliptic problem and $M_c$ is a mass correction term. By carefully choosing $\kappa$, the modified entropy is equivalent to $\|f_\varepsilon - M_c\|^2$ and satisfies: $\frac{d}{dt}E_\varepsilon \lesssim -E_\varepsilon$

3. Macroscopic-Microscopic Decomposition

Utilize decomposition $f_\varepsilon = \langle f_\varepsilon \rangle + f_\varepsilon^\perp$, where:

• $\langle f_\varepsilon \rangle$: velocity average (macroscopic part)
• $f_\varepsilon^\perp$: orthogonal complement (microscopic part)

Handle the macroscopic part through elliptic regularization and control the microscopic part using the dissipativity of operator $L_i$.

Technical Innovations

1. Fine analysis of boundary terms: Through Lemmas 2.1-2.2, transform complex boundary integrals into controllable non-negative terms
2. Unified treatment framework: Simultaneously handle time relaxation ($L_1$) and Fokker-Planck ($L_2$) operators
3. Initial layer correction: Introduce $\psi_\varepsilon$ term in diffusive limits to handle initial mismatch
4. Boundary-adapted relative entropy technique: Successfully extend relative entropy methods to bounded domain settings

Main Results

Theorem 1.1: Long-time Asymptotic Behavior

• Conservative case ($\alpha + \beta = 1$): $\|f_\varepsilon(t) - M_0\| \leq Ce^{-\lambda t}\|f_{in} - M_0\|$
• Non-conservative case ($\alpha + \beta \leq \delta < 1$): $\|f_\varepsilon(t)\| \leq C'e^{-\lambda' t}\|f_{in}\|$

Theorem 1.2: Diffusive Asymptotics

Under Maxwell boundary conditions, as $\varepsilon \to 0$: $\|f_\varepsilon - \rho\|_{L^2([0,T] \times \Omega \times \mathbb{R}^d)} \to 0$

where $\rho$ satisfies the diffusion equation with Neumann boundary conditions: