The project is concerned with solving the problem of identification and qualitative analysis of linear and nonlinear discrete and continuous dynamical systems. This is related to the investigation of new methods for analyzing problems of discretization and numerical solution of systems, analysis of stability, identification of the limit behaviour of delayed systems, analysis of systems with weak after-effect and the application of delayed matrix to solving control problems. The focus will be given to designing estimation procedures for nonlinear models from the Bayesian perspective and adopting a non-iterative learning schema on the basis of global distributional approximation techniques. To deal with an incomplete model of parameter variations, selective forgetting will be considered and adjusted automatically in a way that complies with the degree of the system non-stationarity. The outcomes will include new approaches to the identification and qualitative analysis of nonlinear dynamical systems.

Klíčová slova
Dynamical system; identification; discretization; aftereffect; stability; limit behaviour; control; Bayesian state; nonlinear estimation