| Time: | January 30, 2024 |
|---|---|
| Download as iCal: |
|
Prof. Ivano Notarnicola
Department of Electrical, Electronic, and Information Engineering "Guglielmo
Marconi"
University of Bologna
Bologna, Italy
Tuesday 2024-01-30 4 p.m.
IST Seminar Room 2.255 - Pfaffenwaldring 9 - Campus Stuttgart-Vaihingen
Abstract
In this talk, we will investigate the Alternating Direction Method of Multipliers (ADMM) applied to constraint-coupled optimization problems from a system-theoretic perspective. The algorithmic structure of ADMM can be interpreted as a linear, discrete-time dynamical system in closed loop with a static nonlinearity. By employing tools from Lyapunov theory and passivity, we demonstrate the global asymptotic stability of the origin for the closed-loop system. Equivalently, this implies the global asymptotic convergence of ADMM to the optimal solution of the constraint-coupled optimization problem. The presented approach is then extended to distributed systems, allowing for a systematic design of a decentralized optimization algorithm for the same class of problems.
Biographical Information
Ivano Notarnicola received the MSc in Computer Engineering and the PhD degree in Engineering of Complex Systems from the University of Salento, Lecce (Italy) in 2014 and 2018, respectively. From 2018 to 2020, he was a Post-doctoral fellow with the Department of Electrical Engineering at the University of Bologna (Italy) where he is currently senior assistant professor. He received the 2021 IEEE Transactions on Control of Network Systems Outstanding Paper Award from the IEEE CSS. He is an Associate Editor of the IEEE CSS Conference Editorial Board. His ongoing research focuses on distributed optimization, optimal control algorithms and system theory for optimization algorithms in learning and network systems.