Litcius/Paper detail

Reproducing kernel Hilbert space compactification of unitary evolution groups

Suddhasattwa Das, Dimitrios Giannakis, Joanna Slawinska

2021Applied and Computational Harmonic Analysis53 citationsDOIOpen Access PDF

Abstract

A framework for coherent pattern extraction and prediction of observables of measure-preserving, ergodic dynamical systems with both atomic and continuous spectral components is developed. This framework is based on an approximation of the generator of the system by a compact operator Wτ on a reproducing kernel Hilbert space (RKHS). The operator Wτ is skew-adjoint, and thus can be represented by a projection-valued measure, discrete by compactness, with an associated orthonormal basis of eigenfunctions. These eigenfunctions are ordered in terms of a Dirichlet energy, and provide a notion of coherent observables under the dynamics akin to the Koopman eigenfunctions associated with the atomic part of the spectrum. In addition, Wτ generates a unitary evolution group {etWτ}t∈R on the RKHS, which approximates the unitary Koopman group of the system. We establish convergence results for the spectrum and Borel functional calculus of Wτ as τ→0+, as well as an associated data-driven formulation utilizing time series data. Numerical applications to ergodic systems with atomic and continuous spectra, namely a torus rotation, the Lorenz 63 system, and the Rössler system, are presented.

Topics & Concepts

MathematicsHilbert spaceOrthonormal basisPure mathematicsErgodic theoryEigenfunctionOperator (biology)Reproducing kernel Hilbert spaceObservableTorusMathematical analysisDirichlet formUnitary representationUnitary groupDynamical systems theoryOperator theoryEmbeddingKernel (algebra)Spectrum (functional analysis)Laplace operatorErgodicityDirichlet distributionSpectral theoryUnitary stateGroup (periodic table)Compactification (mathematics)Compact operator on Hilbert spaceDiscrete mathematicsBanach spaceSeries (stratigraphy)Dynamical system (definition)Unitary matrixGeodesicModel Reduction and Neural NetworksControl Systems and IdentificationStability and Controllability of Differential Equations
Reproducing kernel Hilbert space compactification of unitary evolution groups | Litcius