Rigorous data‐driven computation of spectral properties of Koopman operators for dynamical systems

DOI10.1002/CPA.22125arXiv2111.14889OpenAlexW3216122675MaRDI QIDQ6180710FDOQ6180710

Authors: Matthew J. Colbrook, Alex Townsend 

Publication date: 2 January 2024

Published in: Communications on Pure and Applied Mathematics (Search for Journal in Brave)

Abstract: Koopman operators are infinite-dimensional operators that globally linearize nonlinear dynamical systems, making their spectral information valuable for understanding dynamics. However, Koopman operators can have continuous spectra and infinite-dimensional invariant subspaces, making computing their spectral information a considerable challenge. This paper describes data-driven algorithms with rigorous convergence guarantees for computing spectral information of Koopman operators from trajectory data. We introduce residual dynamic mode decomposition (ResDMD), which provides the first scheme for computing the spectra and pseudospectra of general Koopman operators from snapshot data without spectral pollution. Using the resolvent operator and ResDMD, we compute smoothed approximations of spectral measures associated with general measure-preserving dynamical systems. We prove explicit convergence theorems for our algorithms, which can achieve high-order convergence even for chaotic systems when computing the density of the continuous spectrum and the discrete spectrum. Since our algorithms come with error control, ResDMD allows aposteri verification of spectral quantities, Koopman mode decompositions, and learned dictionaries. We demonstrate our algorithms on the tent map, circle rotations, Gauss iterated map, nonlinear pendulum, double pendulum, and Lorenz system. Finally, we provide kernelized variants of our algorithms for dynamical systems with a high-dimensional state space. This allows us to compute the spectral measure associated with the dynamics of a protein molecule with a 20,046-dimensional state space and compute nonlinear Koopman modes with error bounds for turbulent flow past aerofoils with Reynolds number $>10^5$ that has a 295,122-dimensional state space.

Full work available at URL: https://arxiv.org/abs/2111.14889

Recommendations

• Consistent spectral approximation of Koopman operators using resolvent compactification
• Ergodic theory, dynamic mode decomposition, and computation of spectral properties of the Koopman operator
• Data-driven spectral analysis of the Koopman operator
• The mpEDMD Algorithm for Data-Driven Computations of Measure-Preserving Dynamical Systems
• Modern Koopman theory for dynamical systems

zbMATH Keywords

dynamical systemsspectral theorydynamic mode decompositionKoopman operatordata-driven discovery

Mathematics Subject Classification ID

Dynamical aspects of measure-preserving transformations (37A05) Approximation methods and numerical treatment of dynamical systems (37M99) Numerical problems in dynamical systems (65P99)

Cites Work

• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Discovering governing equations from data by sparse identification of nonlinear dynamical systems
• Dynamic mode decomposition of numerical and experimental data
• Spectral analysis of nonlinear flows
• Deterministic Nonperiodic Flow
• Analysis of Fluid Flows via Spectral Properties of the Koopman Operator
• Title not available (Why is that?)
• Title not available (Why is that?)
• Ergodic theory of chaos and strange attractors
• Title not available (Why is that?)
• Title not available (Why is that?)
• Computer-assisted bounds for the rate of decay of correlations
• Harmonic analysis of operators on Hilbert space
• Hydrodynamic stability without eigenvalues
• On the Approximation of Complicated Dynamical Behavior
• Title not available (Why is that?)
• Mathematical Description of Linear Dynamical Systems
• Title not available (Why is that?)
• Title not available (Why is that?)
• A survey of projection-based model reduction methods for parametric dynamical systems
• Chaos, fractional kinetics, and anomalous transport
• The exponentially convergent trapezoidal rule
• Title not available (Why is that?)
• On the Gibbs Phenomenon and Its Resolution
• Title not available (Why is that?)
• CMV matrices: Five years after
• Matrix \(p\)-norms are NP-hard to approximate if \(p\neq1,2,\infty\)
• Title not available (Why is that?)
• Comparison of systems with complex behavior
• Variants of dynamic mode decomposition: boundary condition, Koopman, and Fourier analyses
• Spectral properties of dynamical systems, model reduction and decompositions
• On dynamic mode decomposition: theory and applications
• Applied Koopmanism
• A kernel-based method for data-driven Koopman spectral analysis
• Title not available (Why is that?)
• A data-driven approximation of the koopman operator: extending dynamic mode decomposition
• Discretization of the Frobenius-Perron operator using a sparse Haar tensor basis: the sparse Ulam method
• On the solvability complexity index, the \(n\)-pseudospectrum and approximations of spectra of operators
• On the numerical approximation of the Perron-Frobenius and Koopman operator
• On the level sets of the resolvent norm of a linear operator
• On \((N,\epsilon)\)-pseudospectra of operators on Banach spaces
• Hamiltonian Systems and Transformation in Hilbert Space
• Spectral pollution and how to avoid it
• The rate of convergence in ergodic theorems
• Filters, mollifiers and the computation of the Gibbs phenomenon
• Title not available (Why is that?)
• Family of spectral filters for discontinuous problems
• Title not available (Why is that?)
• Stable reconstructions in Hilbert spaces and the resolution of the Gibbs phenomenon
• Title not available (Why is that?)
• Coherent swing instability of power grids
• Visual complex functions. An introduction with phase portraits
• Data-driven reduced order modeling for time-dependent problems
• Data-driven model reduction and transfer operator approximation
• Data-driven spectral analysis of the Koopman operator
• Delay-coordinate maps and the spectra of Koopman operators
• On the infinite-dimensional QR algorithm
• Data-driven spectral decomposition and forecasting of ergodic dynamical systems
• Data-driven science and engineering. Machine learning, dynamical systems, and control
• Koopman spectra in reproducing kernel Hilbert spaces
• On convergence of extended dynamic mode decomposition to the Koopman operator
• Ergodic theory, dynamic mode decomposition, and computation of spectral properties of the Koopman operator
• Dynamic mode decomposition for financial trading strategies
• The finite section method for TOEPLITZ operators on the quarter-plane with piecewise continuous symbols
• Data-driven approximation of the Koopman generator: model reduction, system identification, and control
• Extended dynamic mode decomposition with dictionary learning: A data-driven adaptive spectral decomposition of the Koopman operator
• Residual dynamic mode decomposition: robust and verified Koopmanism
• Reproducing kernel Hilbert space compactification of unitary evolution groups
• Sparse sensor placement optimization for classification
• Analog forecasting with dynamics-adapted kernels
• Lectures on ergodic theory
• Eigendecompositions of transfer operators in reproducing kernel Hilbert spaces
• Spectrum of the Koopman operator, spectral expansions in functional spaces, and state-space geometry
• Computing spectral measures and spectral types
• Data Driven Modal Decompositions: Analysis and Enhancements
• Koopman analysis of the long-term evolution in a turbulent convection cell
• Foundations of machine learning
• Computing Spectral Measures of Self-Adjoint Operators
• Time-delay observables for Koopman: theory and applications
• Linearly recurrent autoencoder networks for learning dynamics
• Kernel Analog Forecasting: Multiscale Test Problems
• Modern challenges facing turbomachinery aeroacoustics
• Delay-coordinate maps, coherence, and approximate spectra of evolution operators
• Resonant acoustic frequencies of a tandem cascade. Part 1. Zero relative motion
• Koopman resolvent: a Laplace-domain analysis of nonlinear autonomous dynamical systems
• Koopman operator, geometry, and learning of dynamical systems
• The mpEDMD Algorithm for Data-Driven Computations of Measure-Preserving Dynamical Systems

Cited In (8)

• A LAPACK implementation of the dynamic mode decomposition
• Approximation of translation invariant Koopman operators for coupled non-linear systems
• Consistent spectral approximation of Koopman operators using resolvent compactification
• Another look at residual dynamic mode decomposition in the regime of fewer snapshots than dictionary size
• Representing turbulent statistics with partitions of state space. I: Theory and methodology
• EDMD for expanding circle maps and their complex perturbations
• Orthogonal polynomial approximation and extended dynamic mode decomposition in chaos
• A Koopman-Takens theorem: linear least squares prediction of nonlinear time series

This page was built for publication: Rigorous data‐driven computation of spectral properties of Koopman operators for dynamical systems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6180710)