Current modelling and simulation standards for hybrid systems are discussed and an overview of the different structures of neural networks is given. Based on these basic principles, a framework, replacing elements of the hybrid automaton with neural networks, is introduced. It benefits from existing hybrid modelling formalisms by adapting their elements. The framework covers three different cases: the substitution of the dynamical description, the supersession of the discrete process and the replacement of the entire hybrid system. In addition, it enables a generalised structure to define feed forward networks, supporting the modelling process for various application and facilitates interdisciplinary exchange.
Paper
Full text
A Framework Including Artificial Neural Networks in Modelling Hybrid Dynamical Systems
Semantic Scholar · Computer Science · 2020
Abstract
Current modelling and simulation standards for hybrid systems are discussed and an overview of the different structures of neural networks is given. Based on these basic principles, a framework, replacing elements of the hybrid automaton with neural networks, is introduced. It benefits from existing hybrid modelling formalisms by adapting their elements. The framework covers three different cases: the substitution of the dynamical description, the supersession of the discrete process and the replacement of the entire hybrid system. In addition, it enables a generalised structure to define feed forward networks, supporting the modelling process for various application and facilitates interdisciplinary exchange.