2
$\begingroup$

I was reading the paper "Norm inflation for the generalized Navier-Stokes equations" which can be found here: https://arxiv.org/abs/1212.3801.

In Lemma 2.1, the authors said for any $\phi \in L^{\infty}$ and $\alpha>0$ $$\Vert \nabla e^{-t(-\Delta)^{\alpha}}\Vert_{L^\infty} \leq C t^{-\frac{1}{t^{2\alpha}}}\Vert \phi\Vert_{L^\infty} .$$

I am not sure why this is true since the Leray projector $\mathbb P$ is a bounded operator in $L^{2}$. Furthermore, if we want to replace the $L^{\infty}$ norm with some negative Besov norm $\dot B^{-s}_{p,q}$, is the above inequality still true? I appreciate if any reference could be provided.

$\endgroup$

1 Answer 1

1
$\begingroup$

Let $K(x)$ denote the Schwartz kernel of the Fourier multiplier $\nabla e^{-t(-\Delta)^\alpha}$. It's straightforward to check from Fourier inversion and Plancherel's theorem that for $t>0$, $$K(x) = (2\pi)^{-3}\int_{\mathbb{R}^3} i\xi e^{-t|\xi|^{2\alpha}}e^{i\xi\cdot x}d\xi,\qquad\forall x\in\mathbb{R}^3.$$ I claim that $K\in L^1(\mathbb{R}^3)$. To see this, observe that for any multi-index $\beta=(\beta_1,\beta_2,\beta_3)$ of order $|\beta|=4$ and $x\in\mathbb{R}^3$ such that $x^\beta = x_1^{\beta_1}x_2^{\beta_2}x_3^{\beta_3}\neq0$, we have that $$K(x)=\frac{1}{x^\beta (2\pi)^3}\int_{\mathbb{R}^3}i\xi e^{-t|\xi|^{2\alpha}} (\partial_{\xi}^\beta e^{i\xi\cdot x})d\xi.$$ Now integrate by parts in $\xi$. The most singular term in the application of the Leibnitz rule comes from repeatedly differentiating $|\xi|^{2\alpha}$. You can show that $$|K(x)|\lesssim_\alpha \frac{1}{|x^\beta|}\int_{\mathbb{R}^3}|\xi|^{2\alpha-3}e^{-t|\xi|^{2\alpha}}d\xi\lesssim \frac{t^{-1/2\alpha}}{|x^\beta|},$$ where the last expression follows from dilation invariance of Lebesgue measure. Since $|\beta|>3$, the desired conclusion follows.

Regarding negative-order Besov spaces, try playing around with the fact that Littlewood-Paley projectors reduce Fourier multipliers to multiplication by a function of a dyadic frequency.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.