Kinetic Fokker-Planck Equations with Nonlinear Diffusion

Abstract: We study existence, regularity, and uniqueness for the nonlinear kinetic Fokker--Planck equation $$\partial_t f=Δ<em>vΨ(f)-v\cdot\nabla_x f, \qquad f|</em>{t=0}=f_0,$$ on $\mathbb R<sup>{2d}$. In the model case $Ψ(r)=r<sup>s$, this equation couples nonlinear fast-diffusion/porous-medium type diffusion with kinetic transport. A distinctive feature is that the diffusion acts only in the velocity variable $v$, so that compactness in the spatial variable $x$ cannot be obtained from standard elliptic estimates and must instead be recovered through the hypoelliptic structure. Under general structural assumptions on $Ψ$, including the fast-diffusion powers $Ψ(r)=r<sup>s$ with $s\in(0,1)$, we construct nonnegative weak solutions and prove quantitative anisotropic Besov regularity estimates. Under an additional mass-critical growth condition on the fast-diffusion side, the constructed weak solution preserves mass, admits a renormalized kinetic formulation, and is unique in the $L<sup>1$-class of mass-preserving renormalized kinetic solutions. In the power-law case $Ψ(r)=r<sup>s$, this condition is precisely $s\ge 1-1/d$ when $d\ge2$, while in dimension $d=1$ the whole fast-diffusion range $s\in(0,1)$ is covered. The main analytic ingredient is a parameter-dependent smoothing estimate for the kinetic semigroup generated by $$ Ψ&#39;(ζ)Δ_v - v\cdot\nabla_x , $$ which quantitatively tracks the dependence on the kinetic level $ζ$. Combined with the kinetic formulation, this estimate yields compactness in both spatial and velocity variables for the nonlinear hypoelliptic problem. As an application, we also obtain martingale-problem solutions to the associated distributional-density dependent stochastic differential equation.

• The paper introduces a novel kinetic formulation that couples nonlinear, degenerate diffusion in velocity with linear spatial transport.
• It proves the existence, regularity, and uniqueness of weak, renormalized solutions using precise hypoelliptic estimates in anisotropic Besov spaces.
• The study identifies a sharp mass-critical threshold and connects the PDE analysis to a probabilistic martingale framework for mean-field limits.

Analysis of Kinetic Fokker-Planck Equations with Nonlinear Diffusion

Introduction and Model Overview

The paper "Kinetic Fokker-Planck Equations with Nonlinear Diffusion" (2607.00458) investigates the fundamental analytical aspects of a class of nonlinear kinetic equations coupling degenerate, density-dependent diffusion in velocity with linear kinetic transport in phase space. The main object of study is

$$\partial_t f = \Delta_v \Psi(f) - v \cdot \nabla_x f,$$

where $f = f(t, x, v)$, $(x,v) \in \mathbb{R}^{2d}$, and $\Psi$ is a nonlinear function encoding fast diffusion or porous medium-type behavior. This model is paradigmatic as it merges strongly nonlinear degenerate diffusion (via $\Psi$), which acts only on $v$, with spatial transport given by the kinetic drift $-v\cdot\nabla_x f$, thus exhibiting hypoelliptic structure.

Unlike the classical (fully parabolic) nonlinear diffusion equations, this model's diffusion is confined to the velocity component. Consequently, regularity and compactness in $x$ cannot be concluded through standard elliptic or parabolic methods but must leverage the inherent hypoellipticity of the kinetic operator.

The core technical challenge arises from needing to control compactness and regularity in $(x,v)$ with a diffusion acting solely in $v$ and governed by a nonlinear, possibly singular function $f = f(t, x, v)$0.

Main Analytical Results

Framework, Well-posedness, and Regularity

The authors establish a comprehensive well-posedness theory for weak, renormalized kinetic solutions in anisotropic function spaces for a broad class of nonlinearities $f = f(t, x, v)$1 covering classical porous medium ($f = f(t, x, v)$2) and fast diffusion ($f = f(t, x, v)$3) regimes, as well as regularized power-laws.

Their results are predicated on the following structural properties for $f = f(t, x, v)$4:

• $f = f(t, x, v)$5 with positive, continuous derivatives,
• Growth estimates for $f = f(t, x, v)$6 and its derivatives adapted to degenerate and singular behaviors at zero,
• Explicit parameter restrictions ensuring integrability and closing of compactness arguments for arbitrary space dimensions.

The primary analytic contributions can be summarized as:

• Existence: For any admissible initial datum $f = f(t, x, v)$7, there exists a nonnegative weak solution $f = f(t, x, v)$8 satisfying global-in-time energy, mass, and anisotropic Besov regularity bounds. The velocity regularity is quantified through $f = f(t, x, v)$9, where $(x,v) \in \mathbb{R}^{2d}$0 is related to the square root of $(x,v) \in \mathbb{R}^{2d}$1.
• Regularity: The solutions satisfy quantitative regularity in Besov spaces with phase-space adapted anisotropy—specifically, the $(x,v) \in \mathbb{R}^{2d}$2-variables display differing smoothing rates consistent with hypoelliptic structure.
• Uniqueness and Mass Conservation: Under a sharp, dimension-dependent mass-critical growth threshold ($(x,v) \in \mathbb{R}^{2d}$3 for power-type $(x,v) \in \mathbb{R}^{2d}$4), mass is preserved in time, and uniqueness holds in the $(x,v) \in \mathbb{R}^{2d}$5 class of mass-preserving, renormalized kinetic solutions.
• Renormalized Kinetic Formulation: The authors introduce a kinetic formulation encoding the nonlinearity and degeneracy via a family of linear hypoelliptic equations indexed by the kinetic level, along with an associated kinetic defect measure. This formulation facilitates both compactness and uniqueness arguments.
• Quantitative Smoothing for the Parametrized Kinetic Semigroup: A new family of parameter-dependent hypoelliptic estimates is derived for the kinetic semigroup generated by $(x,v) \in \mathbb{R}^{2d}$6, crucially tracking the nonlinear, possibly singular, dependence on $(x,v) \in \mathbb{R}^{2d}$7.

Key Technical Innovations

A notable technical device is the exploitation of the kinetic formulation where, for the indicator $(x,v) \in \mathbb{R}^{2d}$8, one derives

$(x,v) \in \mathbb{R}^{2d}$9

for a suitable nonnegative kinetic measure $\Psi$0. By stacking these linear (in $\Psi$1) problems over the "kinetic variable" $\Psi$2, the authors access the regularization mechanism for the full phase space, albeit with strong parameter dependence due to $\Psi$3. They develop explicit, quantitative anisotropic Besov estimates for the associated semigroups, accommodating the variable and singular structure of $\Psi$4. This allows for strong compactness in both $\Psi$5 and $\Psi$6 via hypoellipticity, circumventing the absence of direct spatial regularization.

A combination of standard a priori velocity estimates, parameter-dependent semigroup smoothing, and kinetic compactness machinery adapted from renormalized conservation law theory yields the existence and compactness framework.

Mass-Critical Threshold and Uniqueness

The sharp threshold $\Psi$7 for the power law nonlinearity emerges in two critical roles:

• As the minimal growth rate for which mass conservation can be rigorously propagated in solutions,
• As the threshold for $\Psi$8-contractivity and uniqueness in the renormalized kinetic framework.

This threshold matches formal scaling computations and is consistent with analogous phenomena in classical porous medium and fast-diffusion equations.

Martingale Problem and Probabilistic Representation

A rigorous characterization of the nonlinear Fokker-Planck equation as the law of a distribution-dependent SDE in phase space is given, via a nonlinear martingale problem for the SDE system

$\Psi$9

where $\Psi$0. The existence of martingale solutions with one-particle marginal given by the PDE solution $\Psi$1 is deduced as an application of mean-field limit arguments within this new analytical setting. The martingale problem formulation connects to the theory of McKean-Vlasov processes with distributional dependence.

Relation to Existing Literature

This work extends and complements several recent studies at the intersection of kinetic theory, nonlinear degenerate PDEs, and probabilistic mean-field limits:

• Previous results on linear or quadratic kinetic Fokker-Planck equations with non-degenerate diffusion in $\Psi$2 are subsumed as special cases, but those approaches do not extend to degenerate, nonlinear cases due to the lack of explicit kernel structure and the nonlinearity.
• Prior work (Brigati et al., 27 Mar 2026, Bouin et al., 30 Jun 2026), focusing on fundamental solutions and entropy–dissipation methods, typically addressed additional friction or confining terms, or assumed more restrictive nonlinearities. The present paper lifts these constraints and develops a theory for the unconfined setting with minimal structural assumptions.
• The kinetic hypoelliptic estimates build on and sharpen the regularity theories developed in [HWZ20, HZZZ24, HRZ25], adapting them to parameter-dependent, nonlinear settings and integrating them with nonlinear compactness approaches from the $\Psi$3-contractivity theory for parabolic conservation laws.
• The critical threshold matches, both formally and in analytic detail, the sharp dichotomy in well-posedness known for fast diffusion equations (cf. [V07, HP85, DK07]), thus demonstrating the optimality and naturality of the obtained results.
• The approach is fully deterministic but provides a platform for further development of stochastic generalizations, e.g., Dean-Kawasaki models, as alluded to in the discussion and recent literature [FG24, FG25, DFG].

Implications and Further Directions

The techniques advanced in this paper supply a robust template for handling kinetic equations with degenerate or singular nonlinear structure. The quantitative Besov regularity and explicit parameter tracking in the kinetic semigroup estimates could be broadly applicable to other nonlinear kinetic models, including spatially inhomogeneous Landau or Boltzmann equations with non-trivial collision integrals, and to stochastic variants where distribution-dependent or fluctuating noise is present.

The martingale problem perspective bridges nonlinear PDE methods with probabilistic propagation of chaos and McKean-Vlasov analysis, facilitating rigorous passage between particle systems with nonlinear interaction and the limiting kinetic PDE, even in highly singular or degenerate regimes.

Potential extensions include:

• Development of the large-deviation theory and fluctuation analysis for the associated particle systems, and derivations of stochastic PDEs of Dean-Kawasaki type,
• Analysis of the long-time asymptotics, invariant measures, and rates of convergence to equilibrium in the absence of confining potentials,
• Extension of the uniqueness and regularity results to inhomogeneous coefficients, more general drift structures, or fully nonlinear stochastic kinetic equations.

Conclusion

This paper succeeds in developing a complete existence, regularity, and uniqueness theory for kinetic Fokker-Planck equations with nonlinear, degenerate diffusion in velocity, utilizing hypoelliptic smoothing and kinetic formulation techniques. The combination of parameter-dependent semigroup estimates, kinetic measure analysis, and renormalized solution frameworks yields a rigorous and optimally sharp well-posedness result valid up to the mass-critical threshold for fast diffusion. These methods provide a flexible analytical basis for future work on nonlinear and stochastic kinetic PDEs and their mean-field limits (2607.00458).