Constructing Neural Network Based Models for Simulating Dynamical Systems

Dynamical systems see widespread use in natural sciences like physics, biology, and chemistry, as well engineering disciplines such circuit analysis, computational fluid dynamics, control. For simple systems, the differential equations governing dynamics can be derived by applying fundamental physical laws. However, for more complex this approach becomes exceedingly difficult. Data-driven modeling is an alternative paradigm that seeks to learn approximation of a system using observations true system. In recent years, there has been increased interest data-driven techniques solve wide range problems physics engineering. This article provides survey different ways construct models dynamical neural networks. addition basic overview, we review related literature outline most significant challenges from numerical simulations must overcome. Based on reviewed identified challenges, provide discussion promising research areas.