Abstract

Set-based methods are rich and versatile tools addressing challenges in analysis and decision problems in engineering related to, for example, safety, stability, robustness, performance for dynamical systems. Nevertheless, although the theory is general enough to apply in almost every meaningful problem formulation, a straightforward implementation in analysis and control design inevitably hits computational complexity barriers. After noting two main methods, we present an alternative approach of computing invariant sets from a computational geometry point of view, changing the shape of polytopic invariant sets and preserve invariance at the same time. The second part of the talk is focused on application-oriented challenges eventually addressed by computing invariant sets and reachability set sequences, related to (i) robotics and autonomous agents (nonlinearities, nonconvex constraints), (ii) scheduling and resource allocation of computing networks (large scale, nonlinear systems) and (iii) cybersecurity (temporal specifications and hybrid systems, state-dependent disturbances).

The seminar will be held in the LR3A , Department of Engineering, and online (zoom): https://newnham.zoom.us/j/92544958528?pwd=YS9PcGRnbXBOcStBdStNb3E0SHN1UT09