MTA Székház, Felolvasóterem 1051 Budapest, Széchenyi István tér 9.
Dynamic models play a key role in many branches of science. In engineering they have a paramount role in model-based simulation, monitoring, control and optimization, as a key-enabler for technological systems to operate reliably and with high performance. The accuracy of the models is crucial for their subsequent use in model-based operations.
While physical laws are an important basis for developing engineering models, developments in sensing technology, wireless communication, and data storage and handling, sometimes referred to as “Big data”, create new opportunities for on-line data-driven modelling on a larger scale than before.
The classical domain of system identification is a subfield of systems and control theory, devoted to the estimation of dynamic models on the basis of measurement data, including the experiment design that provides the appropriate data. System identification has witnessed important developments in handling open-loop and feedback controlled (closed-loop) systems as well as in goal- and control-oriented modelling.
With the growing spatial complexity of engineering systems, e.g., in power networks, transportation networks and industrial production systems, there is a strong need for effective modelling tools for dynamic networks, being considered as interconnected dynamic systems, whose spatial topology may change over time.
In this presentation I will present some principle elements of the system identification theory, including attractive linearly parametrized model structures, the concept of goal- and control-oriented modelling, as well as the development of open-loop and closed-loop identification methods towards the challenge to develop a comprehensive theory for identification in general dynamic networks.