Hardy's inequality for functions vanishing on a part of the boundary (Q2356043): Difference between revisions

Let \(\Omega\) be a bounded domain in \(\mathbb R^d\), \(D\) be a closed part of its boundary \(\partial\Omega\) and let \(p\in(1,\infty)\). The authors study when the Hardy inequality \[ \int_\Omega \Big|\frac{u(x)}{\text{dist}(x,D)}\Big|^p\, \text{d}x \leq c\int_\Omega |\nabla u(x)|^p\, \text{d}x \] holds for all functions \(u\in W^{1,p}_D(\Omega)\) with a constant \(c>0\) independent of \(u\). Here, \(W^{1,p}_D(\Omega)\) denotes the Sobolev space of functions vanishing on \(D\). The general assumptions stated for the inequality to hold are: {\parindent=8mm \begin{itemize}\item[(i)] the set \(D\) is \(\ell\)-thick (in the sense of [\textit{J. Lehrbäck}, Proc. Am. Math. Soc. 136, No. 6, 2193--2200 (2008; Zbl 1151.46021)]) for some \(\ell\in(d-p,d]\), \item [(ii)] the space \(W^{1,p}_D(\Omega)\) can be equivalently normed by the \(L^p(\Omega)\)-norm involving only the gradient of a function \(u\), \item [(iii)] there exists a bounded linear extension operator \(\mathcal E: W^{1,p}_D(\Omega)\to W^{1,p}_D(\mathbb R^d)\). \end{itemize}} The main part of the paper is devoted to finding suitable geometrical conditions guaranteeing the validity of (i)--(iii), and consequently the Hardy inequality mentioned above.