A survey on the resolvent convergence
Joaquim Duran
TL;DR
This survey analyzes how the resolvent of unbounded self-adjoint operators encodes spectral information and provides robust notions of convergence (norm, strong, and weak resolvent convergence) that remain well-defined for operators with varying domains. It systematically relates resolvent convergence to the convergence of spectra and situates these notions within the broader frameworks of strong/weak graph limits and $G$- and $ ext{Γ}$-convergence of quadratic forms, including precise equivalences under compact embedding. A central thread is that norm resolvent convergence yields full spectral convergence (including essential/discrete parts), while strong resolvent convergence may cause spectral contraction, with the Robin Laplacian serving as a concrete exemplar where degenerating boundary conditions lead to Neumann or Dirichlet limits in the norm resolvent sense. The chapter also furnishes a diagram of implications and demonstrates the theory with a detailed elliptic-operator example, highlighting practical pathways to analyze spectral stability under operator convergence. Overall, the work clarifies how resolvent-based frameworks unify spectral convergence with variational and graph-limit approaches in complex Hilbert spaces.
Abstract
This chapter deals with the notion of the resolvent of a self-adjoint operator. We pay special attention to the convergence of unbounded self-adjoint operators in several resolvent senses, and how they are related to the convergence of their spectra. We also explore the relations that these notions of convergence have with the so-called strong graph limit, $G$-convergence, and $Γ$-convergence.
