ProbXiv
Explorersign in
Problem archiveProblem record

Statement

Cheeger's inequality on convex sets

For an open set of finite measure Ω⊂Rn\Omega\subset\mathbb{R}^n, denote λ1(Ω)\lambda_1(\Omega) the first eigenvalue of the Laplacian with Dirichlet boundary conditions, and h(Ω)h(\Omega) the Cheeger constant which may be defined as [h(\Omega)=\inf_{A\subset \Omega}\frac{|\partial A|}{|A|}=\inf_{u\in\mathcal{C}^\infty_c(\Omega)}\frac{\int_{\Omega}|\nabla u|}{\int_{\Omega}u}] Cheeger's inequality is [\lambda_1(\Omega)\geq \frac{1}{4}h(\Omega)^2] and the constant 14\frac{1}{4} is sharp in the sense that the ratio λ1(Ω)h(Ω)2\frac{\lambda_1(\Omega)}{h(\Omega)^2} tends to 14\frac{1}{4} when Ω\Omega is taken to be the unit ball of Rn\mathbb{R}^n for n→+∞n\to +\infty. It is however not sharp when Ω\Omega is restricted to be a planar convex sets, as is shown for instance by the result of I. Ftouhi: [\frac{\lambda_1(\Omega)}{h(\Omega)^2}\geq 0.902] Ilias Ftouhi's conjecture is that the optimal bound is 2π2(2+π)2≈1.387...\frac{2\pi^2}{(2+\sqrt{\pi})^2}\approx 1.387..., which is reached exactly when Ω\Omega is a square.

An optimal set is known to exist, by standard compactness arguments, but it is unknown whether this optimal set is a polygon. A first stepping stone would be to examine the case of polygonal sets, where the Cheeger set (the optimal set ``AA'' in the definition) should coincide with the boundary of Ω\Omega on some part, or leave the boundary tangentially following circle arcs (as described in [kawohl, lachand-robert]).

Record

Added
Links

Comments

Nothing has been published against this problem yet, and nothing has been said about it.

Sign in with an institutional address to take part in the discussion. Reading every thread stays open to everyone.

Sign in

Solve with an agent

Open the statement in a chat, with the problem and the ground rules already written into the prompt.

This opens a third-party site. Nothing is posted back to ProbXiv and nothing you write there is recorded here — what a model gives you is an attempt, which a person still has to check.