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:20230620T143033
UID:402436ee-0f66-11ee-9b72-000e0c3db68b
SUMMARY:Talk of Dr. Romain Postoyan
DESCRIPTION:Dr. Romain Postoyan\nThe Research Center for Automatic Control (CRAN)\nUniversity of Lorraine\nNancy, France \nTuesday 2023-06-27 4 p.m.\nIST Seminar Room 2.255 - Pfaffenwaldring 9 - Campus Stuttgart-Vaihingen \nAbstract\nDynamic programming (DP) plays a major role in various fields including reinforcement learning,\noperations research, computer sciences and of course control engineering. DP allows to solve\ngeneral optimal control problems in terms of dynamical systems and cost functions. In this talk, we\nwill focus on plants described by deterministic nonlinear discrete-time systems whose inputs are\ngenerated by a DP algorithm. Our goal is to establish sufficient conditions under which robust\nstability properties can be established for the closed-loop system. Interestingly, we will see that\nstability can then be exploited to analyse the near-optimality properties of the algorithm used to\ngenerate the control inputs, thereby improving existing near-optimality bounds. We will mainly\nfocus on policy iteration (PI) and we will see that PI actually needs to be modified to be\nwell-posed, in the sense that the optimization problems solved at each iteration are guaranteed to\nadmit a solution. Afterwards, we will briefly see how similar results can be derived for other\nalgorithms (value iteration, optimistic planning). We believe that these results would benefit the\nburgeoning literature on reinforcement learning, where the algorithms are typically assumed to be\nwell-posed without a clear method for verifying it and where robust stability is essential for safe\noperation of the system. \nBiographical Information\nRomain Postoyan received the ``Ingénieur'' degree in Electrical and Control Engineering from\nENSEEIHT (France) in 2005. He obtained the M.Sc. by Research in Control Theory & Application\nfrom Coventry University (United Kingdom) in 2006 and the Ph.D. in Control Theory from Université\nParis-Sud (France) in 2009. In 2010, he was a research assistant at the University of Melbourne\n(Australia). Since 2011, he is a CNRS researcher at CRAN (France). He serves/served as an associate\neditor for the journals: IEEE Transactions on Automatic Control, Automatica, IEEE Control Systems\nLetters and IMA Journal of Mathematical Control and Information. His fields of interest include:\nLyapunov stability, networked control systems, event-triggered control, hybrid systems, dynamic\nprogramming, nonlinear estimation, lithium-ion batteries.
DTSTART;VALUE=DATE:20230627
URL;VALUE=URI:https://www.ist.uni-stuttgart.de/events/Talk-of-Dr.-Romain-Postoyan/
END:VEVENT
END:VCALENDAR