BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:Europe/Stockholm BEGIN:DAYLIGHT DTSTART:20180325T030000 TZOFFSETFROM:+0100 TZOFFSETTO:+0200 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3 TZNAME:CEST END:DAYLIGHT BEGIN:STANDARD DTSTART:20171029T020000 TZOFFSETFROM:+0200 TZOFFSETTO:+0100 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10 TZNAME:CET END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20180503T000000Z UID:42E58F5D-5BC4-4152-B987-A088CC4C08CB DTSTART;TZID=Europe/Stockholm:20180319T101500 DTEND;TZID=Europe/Stockholm:20180319T113000 DESCRIPTION:This talk will address resource allocation in vehicular communi cations. Different from traditional resource allocation\, strong dynamics caused by high mobility in the vehicular environments poses a serious obst acle to the acquisition of high-quality channel state information (CSI). T o deal with the issue\, we investigate the delay impacts of periodic CSI f eedback and develop efficient graph-based centralized resource management schemes to meet the diverse quality-of-service (QoS) requirements in vehic ular networks. To further reduce signaling overhead\, we take advantage of recent advances in reinforcement learning (RL) and develop an effective d istributed resource allocation scheme. We will show that the demanding lat ency and reliability requirements of vehicular communications\, which are hard to model and analyze using traditional methods\, can be explicitly ac counted for in the proposed deep RL framework.\n\nCo-sponsored by: Linköp ing University\n\nSpeaker(s): Prof. Geoffrey Li\, \n\nRoom: Systemet\, flo or 2\, entrance 27\, Bldg: B-building\, \, Linköping University \, Campus Valla\, Linköping\, Ostergotlands lan\, Sweden\, 581 83 LOCATION:Room: Systemet\, floor 2\, entrance 27\, Bldg: B-building\, \, Lin köping University \, Campus Valla\, Linköping\, Ostergotlands lan\, Swed en\, 581 83 ORGANIZER:danyo.danev@liu.se SEQUENCE:8 SUMMARY:2018 IEEE VTS Distinguished Lecture Tour: Resource Allocation in Ve hicular Communication URL;VALUE=URI:https://events.vtools.ieee.org/m/168723 X-ALT-DESC:Description: <br /><p><span lang="EN-US" style="font-family: 'Ca libri'\,sans-serif\; color: black\;">This talk will address resource alloc ation in vehicular communications. Different from traditional resource all ocation\, strong dynamics caused by high mobility in the vehicular environ ments poses a serious obstacle to the acquisition of high-quality channel state information (CSI). To deal with the issue\, we investigate the delay impacts of periodic CSI feedback and develop efficient graph-based centra lized resource management schemes to meet the diverse quality-of-service ( QoS) requirements in vehicular networks. To further reduce signaling overh ead\, we take advantage of recent advances in reinforcement learning (RL) and develop an effective distributed resource allocation scheme. We will s how that the demanding latency and reliability requirements of vehicular c ommunications\, which are hard to model and analyze using traditional meth ods\, can be explicitly accounted for in the proposed deep RL framework.</ span></p> END:VEVENT END:VCALENDAR