Continuity of eigenvalues and shape optimisation for Laplace and Steklov problems

DOI10.1007/S00039-021-00573-5MaRDI QIDQ2232151zbMATH OpenOpenAlexFDO

Authors Alexandre Girouard, Mikhail Karpukhin, Jean Lagacé

Publication date 4 October 2021

Published in Geometric and Functional Analysis. GAFA (Search for Journal in Brave)

Copyright license Creative Commons Attribution 4.0 International

Full work available at URL https://arxiv.org/abs/2004.10784

zbMATH Keywords

Neumann eigenvalues sharp isoperimetric inequalities for Steklov eigenvalues

Mathematics Subject Classification ID

Boundary value problems for second-order elliptic equations (35J25) Estimates of eigenvalues in context of PDEs (35P15) PDEs on manifolds (35R01) Elliptic equations on manifolds, general theory (58J05) Spectral problems; spectral geometry; scattering theory on manifolds (58J50)

Abstract: We associate a sequence of variational eigenvalues to any Radon measure on a compact Riemannian manifold. For particular choices of measures, we recover the Laplace, Steklov and other classical eigenvalue problems. In the first part of the paper we study the properties variational eigenvalues and establish a general continuity result, which shows for a sequence of measures converging in the dual of an appropriate Sobolev space, that the associated eigenvalues converge as well. The second part of the paper is devoted to various applications to shape optimization. The main theme is studying sharp isoperimetric inequalities for Steklov eigenvalues without any assumption on the number of connected components of the boundary. In particular, we solve the isoperimetric problem for each Steklov eigenvalue of planar domains: the best upper bound for the

k

-th perimeter-normalised Steklov eigenvalue is

8 p i k

, which is the best upper bound for the

k

-th area-normalised eigenvalue of the Laplacian on the sphere. The proof involves realising a weighted Neumann problem as a limit of Steklov problems on perforated domains. For

k = 1

, the number of connected boundary components of a maximizing sequence must tend to infinity, and we provide a quantitative lower bound on the number of connected components. A surprising consequence of our analysis is that any maximizing sequence of planar domains with fixed perimeter must collapse to a point.

Review text

In this important paper, the authors continue the work initiated in two other recent papers, [\textit{A. Girouard} et al., Arch. Ration. Mech. Anal. 239, No. 2, 981--1023 (2021; Zbl 1458.35040)] and [\textit{A. Girouard} and \textit{J. Lagacé}, Invent. Math. 226, No. 3, 1011--1056 (2021; Zbl 1477.58019)] about the properties of Steklov eigenvalues, in particular their links with Neumann eigenvalues. Let me give the main results in the Euclidean case, but most of the Theorems are also stated and proved in the case of a compact Riemannian manifold. Let $\Omega$ be a smooth ($C^1$) open set in $\mathbb{R}^d$ and $\beta(x)$ a positive function with good integrability properties. We will denote by $\lambda_k(\Omega,\beta)$ the eigenvalues of the following weighted Neumann problem: $$ \left\lbrace\begin{array}{lc} -\Delta v =\lambda \beta(x) v & \mbox{in } \Omega, \\ \dfrac{\partial v}{\partial n} =0 & \mbox{on } \partial\Omega \\ \end{array}\right. $$ and by $\sigma_k(\Omega)$ the eigenvalues of the Steklov problem $$ \left\lbrace\begin{array}{lc} \Delta u =0 & \mbox{in } \Omega, \\ \dfrac{\partial u}{\partial n} =\sigma u & \mbox{on } \partial\Omega . \\ \end{array}\right. $$ It is natural to normalize these eigenvalues in the following way: $$ \bar{\lambda}_k(\Omega):=\lambda_k(\Omega) Vol(\Omega)^{\frac{2}{d}},\qquad \bar{\sigma}_k(\Omega):=\sigma_k(\Omega) \dfrac{\mathcal{H}^{d-1}(\partial \Omega)}{ Vol(\Omega)^{\frac{d-2}{d}}}. $$ One of the main result in this paper is the following sharp bound for plane domains: $$ \sup_{\Omega\subset \mathbb{R}^2} \bar{\sigma}_k = 8k\pi . $$ The inequality $\bar{\sigma}_k \leq 8k\pi$ was well known, but the sharpness of this inequality was open, even for $k=1$. To prove it, the authors construct a maximizing sequence by removing small disks from $\Omega$ in an appropriate way. They also prove that the number of boundary components of a maximizing sequence (for $k=1$) needs to be unbounded and they provide a quantitative bound of $ \bar{\sigma}_1$ in terms of this number of boundary components. A consequence of their analysis is that any maximizing sequence of planar domains for $\bar{\sigma}_1$ with fixed perimeter must collapse to a point. Considering the link between Steklov and (weighted) Neumann eigenvalues, they prove the following: for any positive function $\beta(x)$ in $L^{d/2}(\Omega)$ ($d\geq 3$) or in $L^1(\log L)^1(\Omega)$ ($d=2$), one can find a sequence $\Omega_\varepsilon$ such that $$ \bar{\sigma}_k(\Omega_\varepsilon) \to \bar{\lambda}_k(\Omega,\beta). $$ (same result for a compact Riemannian manifold). They are also able to find a sequence $\Omega_\varepsilon$ of plane domains such that $$ \bar{\sigma}_1(\Omega_\varepsilon) \to 8\pi \quad \mbox{ and } \bar{\lambda}_k(\Omega_\varepsilon) \to 0 \ \mbox{for each }k. $$ Let me finally mention that this paper contains more interesting results. In particular, the authors consider eigenvalues associated to any Radon measure on a compact Riemannian manifold and they study continuity properties of the eigenvalues for convergence of these measures.

Recommendations

Cites work

Cited in

(32)

This page was built for publication: Continuity of eigenvalues and shape optimisation for Laplace and Steklov problems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2232151)