Modelling and Simulation of Hybrid Systems with Neural Networks

Abstract

Most processes in industry as well as in nature can be rarely described with one simple model. Therefore various modelling methods established in the last years deal with implementations of complex structures. Two of this methods are topic of this contribution. Due to increasing availability of data from multiple resourced research in the field of neural networks increased exponentially. Neural networks are used to imitate the human brain and enable algorithms to make their own decisions. Also in the industrial sector the importance of data increases. Therefore the research field big data is also important in current industrial research projects. Urban infrastructure is one example: Cars driving on their own, gathering information while driving to make decisions on human behalf. But it would be careless to use only data-based models for simulation of complex processes involving heavy machines. Therefore first principle models are still important and the base of modelling and simulation. In this contribution, a comparison of these controversial approaches is discussed.