Dynamical Systems Modeling Using a Neural- Network-Informed Evolutionary Algorithm
Shahab Razavi, Eric R. Gamazon · 2023
A dynamical system is characterized by a deter-ministic rule describing the trajectory of the system. In general, mathematical modeling of such a system from physical principles remains challenging. The algorithmic extraction of the governing equations of dynamical systems from data may address key obstacles. Here, we develop a data-driven methodology for dynamical systems modeling using deep learning and evolutionary computation. The approach may find utility in a variety of scientific applications.