4
$\begingroup$

Let $x \in \mathbb{R}^3$, $t \in [1,\infty)$ and $ u(x,t)$ be a solution of the PDE

$$\partial_{t}^{2}u - \Delta u = 0 \\ u(x,0) = 0 \\ \partial_{t} u(x,0) = v(x)$$

where $v$ and $\partial_{x_i}v$ are both integrable on $\mathbb{R}^3$ for all $1 \leq i \leq 3$.

Show that there exists $C > 0$ such that $|u(x,t)| \leq \frac{C}{t}$

I'm not too sure how to approach this. Before this I just learned about solving some PDEs with Fourier Transforms, but conditions usually involved the initial functions being in Schwartz Space. Any help is appreciated

$\endgroup$
  • 1
    $\begingroup$ what regularity has $u$? $C^2$? what about $v$? $\endgroup$ – supinf Aug 21 '17 at 15:44
7
+50
$\begingroup$

$\DeclareMathOperator{\p}{\partial} $Recall Kirchoff's formula for the solution to the wave equation with initial conditions $u(x,0)=g$ and $\p_tu(x,0)=v(x)$: $$u(x,t)=\frac{1}{4\pi t^2}\int_{\p B(x,t)}g(y) + \nabla g(y)\cdot(y-x) + tv(y)~d\sigma(y).$$ In your case $g=0$ and so the representation formula simplifies significantly. Now notice that for $y\in\p B(x,t)$ that the outward pointing unit normal is $\nu=\frac{y-x}{t}.$ It then follows that \begin{align*} u(x,t) & = \frac{1}{4\pi t^2}\int_{\p B(x,t)}tv(y)\frac{y-x}{t}\cdot\nu ~d\sigma(y) \\ & = \frac{1}{4\pi t^2}\int_{B(x,t)}\operatorname{div}\left(v(y)(y-x)\right) ~d y \\ & = \frac{1}{4\pi t^2}\int_{B(x,t)}3v(y) + \nabla v(y)\cdot(y-x)~d y. \end{align*} Note that we needed to use the divergence theorem to avoid working with a surface integral.

Now for $t \geq 1$ we have \begin{align} |u(x,t)|&\leq\frac{1}{4\pi t^2}\left(3\|v\|_{L^1}+\int_{B(x,t)}\|\nabla v(y)\|\|y-x\|~dy\right)\\ & \leq\frac{1}{4\pi t^2}\left(3\|v\|_{L^1}+t\|\|\nabla v\|\|_{L^1}~dy\right)\\ & \leq \frac{3\|v\|_{L^1}+\|\|\nabla v\|\|_{L^1}}{4\pi t}. \end{align} So by taking $C=\frac{3\|v\|_{L^1}+\|\|\nabla v\|\|_{L^1}}{4\pi}$ you have the desired result. In the above $\|\|\nabla v\|\|_{L^1}$ just means the $L^1$-norm of $\|\nabla v\|$.

$\endgroup$

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.