1
$\begingroup$

Let $\Omega\subset \mathbb{R}^N$ be an open with boundary class $C^1$ and connected. Let, also, $f:\mathbb{R}\to\mathbb{R}$ satisfying $tf(t)\ge 0$ for all $t\in\mathbb{R}$

Show that every $u\in C^2(\Omega)$ satisfying the problem below is constant. Find also $f$ that makes the solution $0$ everywhere

$$\Delta u = f(u) \mbox{ in $\Omega$}\\\frac{\partial u}{\partial \vec n} = 0\mbox{ in $\partial\Omega$}$$

I tried to use the green identity

$$\int_{\Omega}v\Delta u + \nabla v\cdot\nabla u\ d\sigma(y) = \int_{\partial\Omega} u(y)\frac{\partial u}{\partial \vec{n}}d\sigma(y)$$

If we take $v=1$ then $\nabla v = 0$ and since $\frac{\partial u}{\partial \vec n}$ on the boundary:

$$\int_{\Omega} \Delta u\ d\sigma(y) = \int_{\Omega} f(u)\ d\sigma(y) = 0$$

But I don't think I can take anything useful from this integral. I know that $tf(t)\ge 0$ but I don't know how it helps.

I've been thinking about this for a while. It looks like cauchy or neumann problems but I couldn't find anything useful that could help me solving this.

$\endgroup$
  • $\begingroup$ Well, $t f(t)$ might lead naturally to considering $u f(u) = u \Delta u$, so perhaps a different choice of $v$ is in order. $\endgroup$ – user296602 Aug 22 '18 at 3:57
1
$\begingroup$

As indicated in my comment, it's natural to consider $v = u$ instead, because it leads to a natural usage of the fact about $f$. Following Green's identity, we have

$$\int_{\Omega} (u \Delta u + |\nabla u|^2) \, dV = \int_{\partial \Omega} u \frac{\partial u}{\partial \vec n} d\sigma = 0.$$

Now since the first term in the integral above is nonnegative by the assumption on $f$, we can conclude that $\nabla u = 0$ almost everywhere. One can continue from here.

$\endgroup$

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.