Department of Mathematics
Imperial College London

"Dynamic mode decomposition and its variants"

Abstract:

Dynamic mode decomposition (DMD) is a modal factorization and dimensionality reduction technique for data sequences governed by a discrete dynamical system. It produces finite approximations of the eigenvectors of the Koopman operator and is used to extract coherent structures from observables of general evolution processes. We will motivate and present the algorithmic steps of DMD, discuss its extensions and generalizations that have been developed since its introduction and demonstrate its use on a variety of numerical and experimental data from fluid systems.