Let $\,D\subseteq\mathbb{R}^{n-1}$ be a convex bounded domain. Let$A:D\to(0,\infty)$ be a Lipschitz continuous function. Let $\,\Omega\,$ be bounded domain in $\,\mathbb{R}^{n}\,$ of the form \begin{equation} \Omega:=\left\{(x',x_{n})\in\mathbb{R}^{n}|\,x'\in D,\,0\leqslant x_{n}<A(x')\right\}. \end{equation}

Let $u\in C^{1}(\overline{\Omega})$ be such that \begin{equation} u=0\quad\mbox{on}\quad \overline{\Omega} \cap\{x_{n}=0\}. \end{equation} QUESTION: Let $p>1$ and let $\,\mathbb{R}_{+}^{n}=\{x\in\mathbb{R}^{n}|\,x_{n}>0\}.\,$ Given $\,\epsilon>0,\,$ does there exist a function $\,\tilde{u}\in W_{0}^{1,p}(\mathbb{R}_{+}^{n})\,$such that \begin{equation} \tilde{u}|_{\Omega}=u, \mbox{and}\quad \int_{\mathbb{R}_{+}^{n}\setminus\Omega}|\nabla\tilde{u}|^{p}\,dx<\epsilon. \end{equation}

  • $\begingroup$ Suppose Omega is, say, a square or cube, and u(x) = x_n. Can you solve your problem in this or other simplified situation? $\endgroup$ – Tommi Brander Feb 23 '15 at 11:59
  • $\begingroup$ Somehow this reminds me about the definition of $p$-capacity of the set $\Omega$. I suspect the geometry of the set $\Omega$ plays a role here. $\endgroup$ – Juhana Siljander Mar 16 '15 at 20:33

If you want to prove that the answer is no, you can relax the problem, and simplify it. Choose $u=x_n$, $p=2$ and minimise $$ I:=\int_{R^n_{+}} |\nabla v - 1_\Omega e_n|^2 $$ Clearly, if what you are looking for was possible, then there would be a bounded sequence $v_n$ in $W^{1,2}_0$ such that $I(v_n)\leq\frac{1}{n}$, and therefore the minimum would be 0. But you can solve that problem (take $\Omega$ to be a cube or half a disk as suggested in the comments), and find what the minimum is : it is not zero by Liouville's Theorem for example (but you can compute it explicitely for the nice geometries mentioned above).


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.