Koopman operator theory and dynamic mode decomposition in data-driven science and engineering: A comprehensive review

Abstract

Poincaré’s geometric representation, while historically fundamental in dynamical system analysis, faces challenges with high-dimensional and uncertain systems in modern engineering and data analysis. This article extensively explores Koopman Operator Theory (KOT) and Dynamic Mode Decomposition (DMD) within data-driven science and engineering and advocates for a conceptual shift toward observable dynamics, emphasizing KOT’s capacity to capture nonlinear dynamics in infinitedimensional space. The potential practical applications of Koopman-based methods are highlighted. Leveraging Poincaré’s framework, the limitations of traditional methods are discussed. The review also addresses the growing significance of data-driven methodologies for modelling, predicting, and controlling complex systems.