BEGIN:VCALENDAR
VERSION:2.0
PRODID:OpenCms ustutt.16.0.0.32
BEGIN:VTIMEZONE
TZID:Europe/Berlin
X-LIC-LOCATION:Europe/Berlin
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700329T020000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701025T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20230517T145845
UID:8e9e2c66-f4b2-11ed-a3c0-000e0c3db68b
SUMMARY:Vortrag von Prof. Dominic Liao-McPherson
DESCRIPTION:Prof. Dominic Liao-McPherson\nDepartment of Mechanical Engineering \nThe University of British Columbia\nVancouver, Canada\n \nThursday 2023-05-25 10:00 a.m.\nIST Seminar Room 2.255 - Pfaffenwaldring 9 - Campus Stuttgart-Vaihingen \nAbstract\nMany iterative algorithms in optimization, games, and learning can be viewed as dynamical\nsystems with inputs (measurements, historical data, user feedback), internal states (decision\nvariables, state estimates, Lagrange multipliers), outputs (residuals, actuator commands), and\nuncertainties (noise, unknown parameters). The last few years have witnessed a growing interest in\nstudying how learning, optimization and game-theoretic algorithms behave when placed in closed loop\nwith noisy, uncertain, and dynamic physical systems and conversely how systems theory can be\nleveraged to both analyze existing algorithms and synthesize new ones. Notable recent examples\ninclude applying robust control tools (such as integral quadratic constraints) to analyze and\nsynthesize gradient methods, repurposing optimization algorithms as feedback controllers for\nphysical systems, and studying reinforcement learning algorithms using hybrid systems theory.\nThis dynamical systems perspective on algorithms is crucial for tackling some of the challenges\narising when analyzing and synthesizing the algorithms that underlie modern autonomous, AI-driven,\nand socio-technical systems. These include algorithms that: (i) operate online (e.g., running\nalgorithms with streaming data), (ii) include humans in the loop (e.g., autonomous driving,\nrecommender systems), (iii) are interconnected with physical systems (e.g., optimization-based\nfeedback controllers, real-time MPC), (iv) involve self-interested decision makers with shared\nresources (e.g., traffic/logistic networks), or (v) need to be robust to severe uncertainty (e.g.,\nexogenous disturbances, intrinsic randomness, or non-stationarity). These new and timely\napplications are driving the development of new theoretical tools that synergistically build on\ncontrol, optimization, game, and learning theory as well as novel algorithms and computational\ntechniques that are specialized for noisy, time-varying, and dynamic environments.\nIn this talk, I discuss two scenarios where optimization algorithms are used directly as\ncontrollers for physical systems. First, I present Time-distributed Optimization (TDO), a unifying\nframework for studying the system theoretic consequences of computational limits in the context of\nModel Predictive Control (MPC). I show that it is possible to recover the stability and robustness\nproperties of optimal MPC despite limited computational resources and illustrate how a\nsystem-theoretic view of algorithms can be exploited to certify the closed-loop system. Further, I\nillustrate the applicability of the these methods in the real-world through diesel engine, and\nautonomous driving examples. Second, I discuss Feedback Equilibrium Seeking (FES), a design\nframework for dynamic feedback controllers that track solution trajectories of time-varying\ngeneralized equations, such as local minimizers of nonlinear programs or competitive equilibria\n(e.g., Nash) of non-cooperative games. I present tracking error, stability, and robustness results\nfor the sampled-data cased and provide illustrative examples in DC power grids and supply\nchains. \nBiographical Information\nDominic Liao-McPherson obtained his BASc (with High Honours) in Engineering Science (Aerospace\nOption) from the University of Toronto in 2015 and his PhD in Aerospace Engineering and Scientific\nComputing from the University of Michigan in 2020. He was a postdoctoral scholar at the ETH Zürich\nAutomatic Control Lab until the end of 2022 and is now an Assistant Professor at the University of\nBritish Columbia in Vancouver (Canada) where he leads the Algorithms, Optimization, and Control\nLab. His research interests lie at the interface of control systems, optimization, computation and\ngame theory with applications in robotics, energy systems, and manufacturing. \n
DTSTART;VALUE=DATE:20230525
URL;VALUE=URI:https://www.ist.uni-stuttgart.de/de/veranstaltungen/Vortrag-von-Prof.-Dominic-Liao-McPherson/
END:VEVENT
END:VCALENDAR