Sharp stability of some spectral inequalities (Q412038)

Let \(\lambda_i\) denote the Dirichlet eigenvalues and let \(\mu_i\) denote the Neumann eigenvalues of the scalar Laplacian on a Riemannian manifold. The Krahn-Szegő inequality deals with the second Dirichlet eigenvalue and states that \[ \lambda_2(\Omega)\geq\lambda_2(\Theta_2), \] where \(\Theta_2\) is the union of two balls, both having measure \(|\Omega|/2\), with equality holding if and only if \(\Omega=\Theta_2\). The Szegő-Weinberger inequality deals with the first nontrivial eigenvalue \(\mu_2\) of the Neumann spectrum and states that \[ \mu_2(\Omega)\leq\mu_2(B), \] where \(B\) is a ball having the same measure as \(\Omega\), and equality holds if and only if \(\Omega\) is a ball. The authors provide stability estimates for both inequalities by giving quantitative information on how far a certain set is from being the optimal one in terms of how far it is from realizing the given inequalities. The proper notion of the deficit controls a certain suitable concept of asymmetry. The quantitative version of the Szegő-Weinberger inequality \[ \mu_2(B)-\mu_2(\Omega)\geq c_NA(\Omega)^2 \] is sharp. A fundamental difficulty in the analysis arises from the fact that \(\mu_2\) is a multiple eigenvalue. On the other hand, the quantitative version of the Krahn-Szegő inequality which is given in the paper is not optimal.