BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:Turkey BEGIN:DAYLIGHT DTSTART:20380119T061407 TZOFFSETFROM:+0300 TZOFFSETTO:+0300 RRULE:FREQ=YEARLY;BYDAY=3TU;BYMONTH=1 TZNAME:+03 END:DAYLIGHT BEGIN:STANDARD DTSTART:20160907T000000 TZOFFSETFROM:+0300 TZOFFSETTO:+0300 RRULE:FREQ=YEARLY;BYDAY=1WE;BYMONTH=9 TZNAME:+03 END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20191117T195445Z UID:954B63A3-5F8B-440C-A7E5-6CEAB7DC4B3F DTSTART;TZID=Turkey:20191108T124000 DTEND;TZID=Turkey:20191108T143000 DESCRIPTION:08 November 2019 (12:40): IEEE AP/MTT/EMC/ED Turkey Seminar Ser ies (S.57)\n\nSpeaker: Asst. Prof. Naci Saldı\, Özyeğin University\n\nT opic: "Non-signaling approximations of decentralized stochastic control pr oblems"\n\nLocation: Middle East Technical University\, Ankara\, Turkey\n\ nAbstract: Decentralized stochastic control theory studies decisions of ag ents that are acting collectively based on their local information to opti mize a common cost function under stochastic uncertainty. It will be a pro minent avenue of research for many years to come as modern control systems are increasingly decentralized and interconnected. Currently available te chniques (such as classical dynamic programming\, policy iteration\, value iteration\, linear programming\, etc.) in centralized stochastic control does not apply under decentralized and asymmetric information structures. Even for a very simple decentralized stochastic control problem (i.e.\, Wi tsenhausen’s counterexample)\, application of classical methods for obta ining good strategies leads to poor performance. Moreover\, we note that c omputing optimal strategies for decentralized stochastic control problems is in general known to be NP hard. Hence\, we need new viewpoints and new approaches that overcome decentralized nature of the problem. In this talk \, I will consider non-signaling approximation of finite decentralized sto chastic control problems. I will first introduce a hierarchy of control po licies that can be classified in an increasing order as randomized policie s\, quantum-correlated policies\, and non-signaling policies. Then\, I wil l establish an approximation of optimal policies for decentralized stochas tic control systems via extendible nonsignaling policies. I will show that the distance between extendible non-signaling policies and decentralized policies is small if the extension is sufficiently large. Using this resul t\, I will establish a linear programming (LP) approximation of decentrali zed stochastic control problems. Finally\, I will state an open problem re garding computation of optimal value of quantum-correlated policies.\n\nBi o: Naci Saldi received the B.Sc. and M.S. degrees in Electrical and Electr onics Engineering from Bilkent University in 2008 and 2010\, respectively and the Ph.D. degree in Department of Mathematics and Statistics from Quee n’s University in 2015. He was a postdoctoral researcher at the Universi ty of Illinois at Urbana-Champaign before joining the Department of Natura l and Mathematical Sciences at Özyeğin University as an Assistant Profes sor. He is a co-author of the book Finite Approximations in Discrete-Time Stochastic Control\, published by Springer. His research interests include stochastic and decentralized control\, source coding\, mean-field games\, and applied probability.\n\nSpeaker(s): Asst. Prof. Naci Saldı\, \n\nAnk ara\, Ankara\, Türkiye LOCATION:Ankara\, Ankara\, Türkiye ORGANIZER:ozergul@metu.edu.tr SEQUENCE:0 SUMMARY:IEEE AP/MTT/EMC/ED TURKEY CHAPTER SEMINAR SERIES -- SEMINAR 57 URL;VALUE=URI:https://events.vtools.ieee.org/m/211124 X-ALT-DESC:Description: <br /><p><strong>08 November 2019 (12:40): &nbsp\;I EEE AP/MTT/EMC/ED Turkey Seminar Series (S.57)</strong></p>\n<p>Speaker: A sst. Prof. Naci Saldı\, &Ouml\;zyeğin University</p>\n<p>Topic: "Non-sig naling approximations of decentralized stochastic control problems"</p>\n< p>Location:&nbsp\;Middle East Technical University\, Ankara\, Turkey</p>\n <p>Abstract: Decentralized stochastic control theory studies decisions of agents that are acting collectively based on their local information to op timize a common cost function under stochastic uncertainty. It will be a p rominent avenue of research for many years to come as modern control syste ms are increasingly decentralized and interconnected. Currently available techniques (such as classical dynamic programming\, policy iteration\, val ue iteration\, linear programming\, etc.) in centralized stochastic contro l does not apply under decentralized and asymmetric information structures . Even for a very simple decentralized stochastic control problem (i.e.\, Witsenhausen&rsquo\;s counterexample)\, application of classical methods f or obtaining good strategies leads to poor performance. Moreover\, we note that computing optimal strategies for decentralized stochastic control pr oblems is in general known to be NP hard. Hence\, we need new viewpoints a nd new approaches that overcome decentralized nature of the problem. In th is talk\, I will consider non-signaling approximation of finite decentrali zed stochastic control problems. I will first introduce a hierarchy of con trol policies that can be classified in an increasing order as randomized policies\, quantum-correlated policies\, and non-signaling policies. Then\ , I will establish an approximation of optimal policies for decentralized stochastic control systems via extendible nonsignaling policies. I will sh ow that the distance between extendible non-signaling policies and decentr alized policies is small if the extension is sufficiently large. Using thi s result\, I will establish a linear programming (LP) approximation of dec entralized stochastic control problems. Finally\, I will state an open pro blem regarding computation of optimal value of quantum-correlated policies .</p>\n<p>Bio: Naci Saldi received the B.Sc. and M.S. degrees in Electrica l and Electronics Engineering from Bilkent University in 2008 and 2010\, r espectively and the Ph.D. degree in Department of Mathematics and Statisti cs from Queen&rsquo\;s University in 2015. He was a postdoctoral researche r at the University of Illinois at Urbana-Champaign before joining the Dep artment of Natural and Mathematical Sciences at &Ouml\;zyeğin University as an Assistant Professor. He is a co-author of the book Finite Approximat ions in Discrete-Time Stochastic Control\, published by Springer. His rese arch interests include stochastic and decentralized control\, source codin g\, mean-field games\, and applied probability.</p> END:VEVENT END:VCALENDAR