1
$\begingroup$

Let $\Omega \subset R^n$ a open set and $p >1$. Let $K \subset \Omega$ a compact set. Define $$A:= \{ u \in C^{\infty}_{0} (\Omega) ; \ u \geq 1 \text{ on } K\}$$

$$ B:= \{ u \in C^{\infty}_{0} (\Omega) ; \ u \geq 1 \text{ on } \partial K\}$$

Show that

$$ \inf_{u \in A} \int_{\Omega} |\nabla u |^p \ dx = \inf_{u \in B} \int_{\Omega} |\nabla u |^p \ dx$$

A good time ago i tried to prove this and i only proved the obvious inequality (first inf $\geq$ second inf)...

Someone can give me a hint ?

$\endgroup$
  • 1
    $\begingroup$ Equivalence of two definitions of capacity would be a far better title. Think of this question being in the archive alongside hundreds of questions about infima and suprema in real analysis. $\endgroup$ – 40 votes Jul 16 '13 at 21:00
2
$\begingroup$

Given $u\in B$, observe that $u$ is Lipschitz on an open set containing $K$. So is $\tilde u=\max(u,1)$, and this truncation does not increase the norm of gradient. Next, let $v=1$ on the interior of $K$, and $v= \tilde u$ otherwise. The Lipschitz condition survives. Therefore, $v\in W^{1,p}_0(\Omega)$. Let $\psi $ be a compactly supported smooth function with small $W^{1,p}$ norm such that $\psi >0$ on $K$. Mollifying $v+\psi$, we get a function $w\in C_0^\infty$ such that $w>1$ on $K$ and $\|w-\tilde u\|_{W^{1,p}}$ is small. Since $w\in A$, and $\int|\nabla w|^p$ is not much larger than $\int|\nabla u|^p$, the statement follows.

$\endgroup$
  • $\begingroup$ can you say to me a good book about capacity theory? (my english is terrible, sorry..) $\endgroup$ – math student Jul 16 '13 at 21:46
  • $\begingroup$ @LeandroTavares There are different approaches to capacity: via energy of measures and via Dirichlet integral, as in your question. The energy approach, see Landkof or Armitage-Gardiner. For the Dirichlet integral approach, see Heinonen-Kilpelainen-Martio. Be advised that these books are not exclusively about capacity; it is merely a tool for doing something else. $\endgroup$ – 40 votes Jul 16 '13 at 22:17
  • $\begingroup$ @LeandroTavares One more: section 4.7 in Measure Theory and Fine Properties of Functions by Evans and Gariepy. It's not long, but it covers the main points. $\endgroup$ – 40 votes Jul 18 '13 at 1:15

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.