On the best constant for Hardy’s inequality in $\mathbb {R}^n$ (Q4211099)

Let \(\Omega\) be a domain in \(\mathbb{R}^n\) with a non-empty boundary and \(\delta(x):= \text{dist}(x, \partial\Omega)\), \(x\in\mathbb{R}^n\). For \(p\in (1,\infty)\), the Hardy inequality \[ \Biggl(\int_\Omega| u(x)/\delta(x)|^p dx\Biggr)^{1/p}\leq \gamma\Biggl(\int_\Omega|\nabla u(x)|^p dx\Biggr)^{1/p},\quad u\in\mathring W_{1,p}(\Omega),\tag{1} \] holds for a large family of domains \(\Omega\) including all bounded domains with Lipschitz boundary. The best possible constant \(\gamma= \gamma_p(\Omega)\) in the inequality (1) satisfies \[ \Biggl({1\over\gamma}\Biggr)^p= \mu_p(\Omega):= \inf_{u\in\mathring W_{1,p}(\Omega)} \Biggl(\int_\Omega|\nabla u|^p\Biggl/ \int_\Omega| u/\delta|^p\Biggr).\tag{2} \] The aim of the paper is to clarify the connection between the value of \(\mu_p(\Omega)\) and the existence of a minimizer of the Rayleigh quotient appearing on the right-hand side of (2). The main results state that (i) \(\mu_p(\Omega)\leq c_p:= (1-1/p)^p\); the equality holds here if problem (2) has no minimizer; (ii) \(\mu_2(\Omega)= c_2\) if and only if problem (2) has no minimizer; (iii) \(\mu_2(\Omega)< c_2= 1/4\) if and only if the Rayleigh quotient possesses a minimizer.