An optimal anisotropic Poincaré inequality for convex domains (Q453225)

A classical result of \textit{L. E. Payne} and \textit{H. F. Weinberger} [Arch. Ration. Mech. Anal. 5, 286--292 (1960; Zbl 0099.08402)] gives the optimal constant for Poincaré`s inequality on a convex domain \(\Omega\) in \(n\)-space as \(\pi^2/d^2\), where \(d\) is the diameter of \(\Omega\). In the corresponding Rayleigh-quotient the Dirichlet integral is given in terms of the Euclidean norm of the gradient of the underlying eigenfunction. If this norm is replaced by a more general convex one \(F\), we speak of the anisotropic situation, in which de definition of diameter and convexity has a natural extension. Now the diameter of a domain \(\Omega\) is determined by the smallest anisotropic ball in the dual norm \(F_0\) which contains \(\Omega\). The paper contains also a similar lower bound of the first eigenvalue of the anisotropic Dirichlet-Laplacian. (There is a typo in the statement of Theorem 1.2: \(\lambda_1^N\) should be \(\lambda_1^D\).) Another lower bound for the anisotropic Dirichlet Laplacian (in terms of the so-called Cheeger constants) was given by \textit{M. Novaga} and the reviewer [J. Convex Anal. 15, No. 3, 623--634 (2008; Zbl 1186.35115)].