scientific article; zbMATH DE number 16001

From MaRDI portal
Publication:3972648





[For the entire collection see Zbl 0713.00009.] Let \(\Omega\) be a bounded domain in \(\mathbb{R}^ 2\) and write \(M\) for the unit ball in \(H_ 0^ 1(\Omega)\) using the Dirichlet norm. It is known that \(\sup_{u\in M}\int_ \Omega\exp(\alpha u^ 2)dx\) is finite if and only if \(\alpha\leq 4\pi\). Moreover the supremum is attained (when it is finite) by some \(u_ \alpha\in M\). The higher-dimensional analog, in which \(\exp(\alpha u^ 2)\) is replaced by \(u^ p\), is quite different because the corresponding supremum is finite if and only if \(p\leq 2n/(n- 2)\) but this supremum is only attained for \(p<2n/(n-2)\). Morover when \(\Omega\) is the unit ball, \(u_ p(0)\) becomes infinite as \(p\uparrow 2n/(n-2)\). The authors study the behavior of \(u_ \alpha\) (in two dimensions) as \(\alpha\uparrow 4\pi\) when \(\Omega\) is the unit ball. With a suitable normalization, they announce that \((u_ \alpha)\) converges uniformly on compact subsets of \(\overline\Omega-\{0\}\) and that \(u_ \alpha(0)\to u_ 4\pi(0)\). Proofs will be in a forthcoming paper of the authors [Arch. Ration. Mech. Anal. 106, No. 3, 261-285 (1989; Zbl 0687.46017)].











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3972648)