Sparse Identification of Nonlinear Dynamics for Delay and Stochastic Differential Equations
Please login to view abstract download link
Recently, data-driven model discovery has emerged has a powerful approach to recover governing equations of dynamical systems from temporal data series [3]. In particular, the SINDy algorithm, initially proposed for learning the right-hand side of ordinary differential equations [4], has been extended and applied to diverse classes of problems, including delay differential equations [2, 5] and stochastic (ordinary) differential equations [1]. In this talk we present a further development by proposing a new SINDy algorithm to address the case of stochastic delay differential equations. The relevant MATLAB implementation is tested on several examples, including stochastic models with delay used to describe and investigate supply chains.
