Spectral gaps and approximate invariance in Hilbert spaces

Abstract

This paper develops a quantitative theory of approximate invariant subspaces for bounded linear operators on complex Hilbert spaces. We provide novel characterizations for the existence of \emph{almost invariant} subspaces of nonself-adjoint operators using spectral gap conditions, and we investigate how the geometry of the spectrum influences near-invariance. In particular, we extend Lomonosov's Invariant Subspace Theorem to approximate settings by leveraging perturbation theory and compactness arguments. Applications to quantum decoherence are explored, highlighting the role of approximate invariance in the emergence of classical behavior from quantum systems. Additionally, we propose new criteria for approximate spectral clustering and the stability of invariant subspaces under small perturbations, providing a robust framework for analyzing operator dynamics in infinite-dimensional spaces. Our results contribute original tools and insights for the spectral analysis of operators beyond the self-adjoint and compact settings.