Estimates of variational eigenvalues on metric measure spaces
Prashanta Garain, Alexander Ukhlov
TL;DR
This work extends the theory of Neumann variational eigenvalues to general doubling metric measure spaces, addressing nonlinear $(p,q)$-Laplacian-type operators. It formulates the first nontrivial variational Neumann $(p,q)$-eigenvalue as a minimization over the Newtonian-Sobolev space $N^{1,p}(\Omega)$ with a zero-$L^q$-mean constraint and proves existence of minimizers under Sobolev $(p,p^*)$-embedding domain assumptions. A key insight is identifying $\lambda_{p,q}(\Omega)^{-1/p}$ with the best constant in the $(p,q)$-Sobolev-Poincaré inequality, and the work leverages extension operators and compact embeddings to establish the variational framework. The resulting eigenvalue estimates relate spectral data to extension norms, enabling comparisons across nested domains and generalizing Euclidean results to metric-measure spaces.
Abstract
In the article, we study variational eigenvalues on doubling metric measure spaces. We prove existence of minimizers of variational Neumann $(p,q)$-eigenvalues on metric measure spaces and on this base we obtain estimates of Neumann eigenvalues.
