[Paper Review] Kinetic maximal $L^2$-regularity for the (fractional) Kolmogorov equation

Lukas Niebel, Rico Zacher|arXiv (Cornell University)|Jun 20, 2020

Advanced Mathematical Physics Problems32 references7 citations

TL;DR

This paper establishes kinetic maximal $L^2$-regularity with temporal weights for the fractional Kolmogorov equation, characterizing solution regularity in anisotropic Sobolev spaces. It proves that solutions gain $\frac{\beta}{\beta+1}$ derivatives in position $x$ and achieve instantaneous $C^\infty$ smoothing in $(0,T)\times\mathbb{R}^{2n}$, even for initial data with limited regularity.

ABSTRACT

We introduce the notion of kinetic maximal $L^2$-regularity with temporal weights for the (fractional) Kolmogorov equation. In particular, we determine the function spaces for the inhomogeneity and the initial value which characterize the regularity of solutions to the fractional Kolmogorov equation in terms of fractional anisotropic Sobolev spaces. It is shown that solutions of the homogeneous (fractional) Kolmogorov equation define a semi-flow in a suitable function space and the property of instantaneous regularization is investigated.

Motivation & Objective

• To establish a precise $L^2$-regularity theory for the (fractional) Kolmogorov equation in the presence of temporal weights.
• To characterize the function spaces for the inhomogeneity and initial data that ensure maximal $L^2$-regularity in anisotropic Sobolev spaces.
• To investigate the phenomenon of instantaneous regularization in solutions, particularly the transfer of regularity from velocity $v$ to position $x$.
• To extend known regularity transfer results (e.g., Bouchut's theorem) to the weighted $L^2$-setting with temporal weights.
• To prove that solutions to the homogeneous equation become smooth in space and time immediately after $t=0$, even for rough initial data.

Proposed method

• Introduce a new notion of kinetic maximal $L^2$-regularity with temporal weights $t^{1-\mu}$ for $\mu \in (\frac{1}{2},1]$.
• Define solution spaces $\mathcal{T}_\mu((0,T);X)$ and $L^2_\mu((0,T);X)$ to incorporate time-weighted estimates for the time derivative and velocity derivatives.
• Use the isomorphism $\Phi_\mu(u)(t) = t^{1-\mu}u(t)$ to transfer known $L^2$-estimates from unweighted to weighted settings.
• Apply the theory of semigroups and spectral multipliers to analyze the Kolmogorov operator $\partial_t + v\cdot\nabla_x + (-\Delta_v)^{\beta/2}$.
• Establish a gain of $\frac{\beta^2}{2(\beta+1)}$ additional regularity in $x$ by combining weighted estimates with interpolation and equicontinuity arguments.
• Iterate regularity gains through successive time shifts to prove $C^\infty$ smoothing in $(0,T)\times\mathbb{R}^{2n}$.

Experimental results

Research questions

• RQ1 What function spaces for the inhomogeneity and initial data ensure maximal $L^2$-regularity for the fractional Kolmogorov equation with temporal weights?
• RQ2 How much regularity in the position variable $x$ can be gained from the velocity regularity in the fractional Kolmogorov equation?
• RQ3 Can the solution to the homogeneous fractional Kolmogorov equation achieve $C^\infty$ regularity immediately after $t=0$?
• RQ4 Is the regularity transfer from $v$ to $x$ reflected in the initial data's regularity in $x$?
• RQ5 How do temporal weights improve the regularity estimates and enable the proof of instantaneous smoothing?

Key findings

• The solution space $\mathcal{T}_\mu((0,T);L^2)\cap L^2_\mu((0,T);X_\beta^1)\cap L^2_{1-\beta}((0,T);\dot{H}_x^\beta)$ embeds continuously into $C((0,T];H^{\beta/2}(\mathbb{R}^{2n}))$.
• Solutions gain $\frac{\beta}{\beta+1}$ derivatives in the position variable $x$, consistent with known results, but now in a global, weighted $L^2$-setting.
• The solution to the homogeneous equation belongs to $C^\infty((0,T)\times\mathbb{R}^{2n})$ for any $T>0$, even if the initial data has limited regularity.
• The gain of regularity in $x$ is equivalent to the regularity of the initial data in $x$, while regularity in $v$ corresponds to joint regularity in $x$ and $v$ of the initial data.
• The temporal weight $t^{1-\mu}$ with $\mu > \frac{1}{2}$ enables the proof of $C^\infty$ smoothing via iterative regularity gains through time shifts.
• The result extends Bouchut’s theorem to the weighted $L^2$-framework, showing $u \in L^2_\mu((0,T);H_x^{r/(r+1)})$ for $u \in \mathcal{T}_\mu((0,T);L^2)\cap L^2_\mu((0,T);H_v^r)$.