English

Mean curvature bounds and eigenvalues of Robin Laplacians

Spectral Theory 2015-10-02 v1 Analysis of PDEs Differential Geometry Optimization and Control

Abstract

We consider the Laplacian with attractive Robin boundary conditions, QαΩu=Δu,un=αu on Ω, Q^\Omega_\alpha u=-\Delta u, \quad \dfrac{\partial u}{\partial n}=\alpha u \text{ on } \partial\Omega, in a class of bounded smooth domains ΩRν\Omega\in\mathbb{R}^\nu; here nn is the outward unit normal and α>0\alpha>0 is a constant. We show that for each jNj\in\mathbb{N} and α+\alpha\to+\infty, the jjth eigenvalue Ej(QαΩ)E_j(Q^\Omega_\alpha) has the asymptotics Ej(QαΩ)=α2(ν1)Hmax(Ω)α+O(α2/3), E_j(Q^\Omega_\alpha)=-\alpha^2 -(\nu-1)H_\mathrm{max}(\Omega)\,\alpha+{\mathcal O}(\alpha^{2/3}), where Hmax(Ω)H_\mathrm{max}(\Omega) is the maximum mean curvature at Ω\partial \Omega. The discussion of the reverse Faber-Krahn inequality gives rise to a new geometric problem concerning the minimization of HmaxH_\mathrm{max}. In particular, we show that the ball is the strict minimizer of HmaxH_\mathrm{max} among the smooth star-shaped domains of a given volume, which leads to the following result: if BB is a ball and Ω\Omega is any other star-shaped smooth domain of the same volume, then for any fixed jNj\in\mathbb{N} we have Ej(QαB)>Ej(QαΩ)E_j(Q^B_\alpha)>E_j(Q^\Omega_\alpha) for large α\alpha. An open question concerning a larger class of domains is formulated.

Keywords

Cite

@article{arxiv.1407.3087,
  title  = {Mean curvature bounds and eigenvalues of Robin Laplacians},
  author = {Konstantin Pankrashkin and Nicolas Popoff},
  journal= {arXiv preprint arXiv:1407.3087},
  year   = {2015}
}

Comments

15 pages