Abstract
Stochastic control problems for switching dynamical systems are of paramount importance in control theory. Despite significant progress in recent years, many open challenges remain, particularly in settings with partial observations. This talk will delve into specific research directions concerning the control and filtering of Markov Jump Linear Systems (MJLSs). Key topics under discussion will include the characterization of induced system norms for MJLSs, open problems in optimal linear filtering, and the treatment of detector-based control strategies by means of the homogenization analysis of two-time-scale Markov switching systems.
Biography
Marcos Todorov received the D.Sc. degree in Computational Modeling from the National Laboratory of Scientific Computing (LNCC), Brazil, in 2011. Since 2013, he has been with the Department of Mathematical and Computational Methods, LNCC, where he is currently an associate researcher. He co-authored the book Continuous-Time Markov Jump Linear Systems (Springer, 2013). His research interests include stochastic control, filtering, optimization, and applications of Markov jump linear systems.
