A Data-Driven Approximation of the Koopman Operator: Extending Dynamic Mode Decomposition
read the original abstract
The Koopman operator is a linear but infinite dimensional operator that governs the evolution of scalar observables defined on the state space of an autonomous dynamical system, and is a powerful tool for the analysis and decomposition of nonlinear dynamical systems. In this manuscript, we present a data driven method for approximating the leading eigenvalues, eigenfunctions, and modes of the Koopman operator. The method requires a data set of snapshot pairs and a dictionary of scalar observables, but does not require explicit governing equations or interaction with a "black box" integrator. We will show that this approach is, in effect, an extension of Dynamic Mode Decomposition (DMD), which has been used to approximate the Koopman eigenvalues and modes. Furthermore, if the data provided to the method are generated by a Markov process instead of a deterministic dynamical system, the algorithm approximates the eigenfunctions of the Kolmogorov backward equation, which could be considered as the "stochastic Koopman operator" [1]. Finally, four illustrative examples are presented: two that highlight the quantitative performance of the method when presented with either deterministic or stochastic data, and two that show potential applications of the Koopman eigenfunctions.
This paper has not been read by Pith yet.
Forward citations
Cited by 13 Pith papers
-
Diagnostic Certificates of Data Quality and Regression Identifiability for Koopman Identification
The paper introduces diagnostic certificates that separately assess state-space coverage, lifted-feature nondegeneracy, and regression-spectrum quality for Koopman and EDMDc identification, with theoretical guarantees...
-
Controlling Transient Amplification Improves Long-horizon Rollouts
Commutativity regularization on Jacobians reduces transient error amplification in neural simulators, enabling stable rollouts over thousands of steps on physical and climate data.
-
Deep-Koopman-KANDy: Dictionary Discovery for Deep-Koopman Operators with Kolmogorov-Arnold Networks for Dynamics
Deep-Koopman-KANDy recovers symbolic Koopman dictionaries post-training by replacing the encoder and decoder with KANs and applying a level-set construction with chain-rule gradients, achieving high recall on Lorenz a...
-
Data-driven discovery of polynomial ODEs with provably bounded solutions
SILAS jointly optimizes polynomial ODE vector fields and polynomial Lyapunov functions from data to produce models with provably bounded trajectories via compact absorbing sets.
-
Conditional Score-Based Modeling of Effective Langevin Dynamics
A novel identity connects reduced-model drift and diffusion to the conditional score of the finite-time transition density, turning calibration into a least-squares problem over stationary lagged pairs that preserves ...
-
Data-driven analysis of metastability in a stochastic bistable system
Data-driven Koopman analysis of a bistable stochastic system recovers large deviation theory escape time statistics and basin structure via the subdominant mode.
-
Controlling Transient Amplification Improves Long-horizon Rollouts
Commutativity regularization mitigates transient error amplification in autoregressive neural simulators by penalizing non-normality and non-commutativity of Jacobians, yielding stable long-horizon rollouts.
-
Numerical approximation of the Koopman-von Neumann equation: Operator learning and quantum computing
Data-driven approximation methods are derived for the unitary Koopman-von Neumann operator, its eigenvalues and eigenfunctions, with explicit quantum-circuit representations for finite-dimensional projections.
-
Learning graphons from data: Random walks, transfer operators, and spectral clustering
Connects continuum stochastic signals to graphon random walks via Koopman and Perron-Frobenius operators for spectral clustering and graphon reconstruction from data.
-
Optimization of randomized neural networks for transfer operator approximation
Optimizing the activation function in randomized neural networks provides a more suitable dictionary for transfer operator approximation in stochastic differential equations and random walks on graphons.
-
Koopman Representations for Early Outbreak Warning and Minimal Counterfactual Intervention in Multi-Agent Epidemic Simulations
Koopman latent space representations from early epidemic simulation data enable accurate prediction of major outbreaks and identification of minimal single-agent interventions to prevent them.
-
Spatio-Temporal Prediction via Operator-Valued RKHS and Koopman Approximation
A theoretical framework establishing representer theorems, Sobolev approximation bounds, and spectral convergence for kernel-based learning of spatio-temporal dynamical systems using OV RKHS and Koopman approximations.
-
Latent Linear Quadratic Regulator for Robotic Control Tasks
LaLQR learns a latent linear-quadratic representation of robotic systems by imitating MPC to enable efficient LQR control.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.