Let $d\geq 2$, $I\subset\mathbb{R}$ a time interval and $t_0\in I$. Let $u,v$ be solutions to the free Schrodinger equation on $\mathbb{R}^d$, i.e. $(i\partial_t+\Delta)u=(i\partial_t+\Delta)v=0.$ Let us fix $j,k\in\mathbb{Z}$. For every $\delta\geq 0$, we have the following Bourgain-type bilinear estimate

$$\Vert (P_ju)(P_kv)\Vert_{L^2(I\times\mathbb{R}^d)}\leq C(\delta)\Vert P_ju(t_0)\Vert_{\dot{H}^{-1/2+\delta}}\Vert P_kv(t_0)\Vert_{\dot{H}^{\frac{d-1}{2}-\delta}},$$ where $P_j$, $P_k$ are the Paley-Littlewood projections.

My doubt is the following: Can we choose the constant $C(\delta)$ to be uniform in $j,k$?

  • $\begingroup$ I think your norm $\dot{H}^{1-\delta}$ should instead be $\dot{H}^{\frac{d-1}{2}-\delta}$. $\endgroup$ – Matt Rosenzweig Jul 20 at 4:04
  • $\begingroup$ Sure, I was thinking in three dimension. I edited $\endgroup$ – Capublanca Jul 22 at 18:35

I don't understand why you have the $\delta$. One has the following estimate (see, for instance, Theorem 4.18 in the Clay lecture notes "Nonlinear Schrodinger Equations at Critical Regularity" by Rowan Killip and Monica Visan):

Bilinear Strichartz. Let $j,k$ be integers such that $j\leq k$. Then $$\|(e^{it\Delta}P_j f)(e^{it\Delta}P_k g)\|_{L_{t,x}^2(\mathbb{R}\times\mathbb{R}^d)} \lesssim 2^{\frac{j(d-1)}{2} -\frac{k}{2}} \|f\|_{L^2(\mathbb{R}^d)}\|g\|_{L^2(\mathbb{R}^d)}.$$

Let$\tilde{P}_j$ and $\tilde{P}_k$ be "fattened" Littlewood-Paley projectors such that their symbols are identically one on the support of the symbol of $P_j$ and $P_k$, respectively, for all $j,k\in\mathbb{Z}$, so that in particular, $P_j\tilde{P}_j = P_j$ and $P_k\tilde{P}_k=P_k$. We then have as a consequence of Bilinear Strichartz that

$$\|(e^{it\Delta}P_j f)(e^{it\Delta}P_k g)\|_{L_{t,x}^2(\mathbb{R}\times\mathbb{R}^d)} \lesssim 2^{\frac{j(d-1)}{2}-\frac{k}{2}} \|\tilde{P}_j f\|_{L^2(\mathbb{R}^d)} \|\tilde{P}_k g\|_{L^2(\mathbb{R}^d)}.$$ By Bernstein's lemma, \begin{align} 2^{\frac{j(d-1)}{2}}\|\tilde{P}_j f\|_{L^2(\mathbb{R}^d)} &\sim \|\tilde{P}_j f\|_{\dot{H}^{\frac{d-1}{2}}(\mathbb{R}^d)} \lesssim\|f\|_{\dot{H}^{\frac{d-1}{2}}(\mathbb{R}^d)},\\ 2^{-\frac{k}{2}}\|\tilde{P}_k g\|_{L^2(\mathbb{R}^d)} &\sim \|\tilde{P}_kg\|_{\dot{H}^{-1/2}(\mathbb{R}^d)} \lesssim \|g\|_{\dot{H}^{-1/2}(\mathbb{R}^d)}. \end{align}

  • $\begingroup$ Yeah your are completely right, I was getting confusing by the fact that for non-localised data u need the delta, but here is completely fine. $\endgroup$ – Capublanca Jul 22 at 18:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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