BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//UNC Statistics & Operational Research - ECPv5.16.0//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:UNC Statistics & Operational Research
X-ORIGINAL-URL:https://stor.unc.edu
X-WR-CALDESC:Events for UNC Statistics & Operational Research
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20200308T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20201101T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20200306T153000
DTEND;TZID=America/New_York:20200306T163000
DTSTAMP:20220811T175242
CREATED:20200302T111919Z
LAST-MODIFIED:20200302T111919Z
UID:9920-1583508600-1583512200@stor.unc.edu
SUMMARY:Graduate Seminar: Alex Touzov
DESCRIPTION:Graduate Seminar \nFriday\, March 6th\, 2020 \n120 Hanes Hall \n3:30pm \n \nAlex Touzov \nUNC-Chapel Hill \n \nWeak infeasibility in semidefinite programming: a complete characterization and generating all instances \n \nAbstract: In this work\, we analyze weakly infeasible semidefinite programs (SDPs). These SDPs are infeasible\, but an arbitrarily small perturbation can make them feasible. \n \nWeakly infeasible SDPs appear in many guises\, some classical and some more recent: 1) as asymptotes of the semidefinite cone; as 2) difficult SDPs\, which are often mistaken for feasible ones by even the best solvers; and as 3) infeasible SDPs within zero distance to ill-posedness. \n \nWe first describe a very simple combinatorial characterization of weak infeasibility in SDPs. The characterization uses elementary operations (inherited from Gaussian elimination) and reformulates the original SDP into a form that makes the weak infeasibility trivial to see. We then introduce two simple combinatorial algorithms to generate any weakly infeasible SDP: with a suitable starting data\, any such SDP is obtained as an output of our algorithms. We conclude with a computational study. \n
URL:https://stor.unc.edu/event/graduate-seminar-alex-touzov/
LOCATION:120 Hanes Hall\, Hanes Hall\, Chapel Hill\, NC\, 27599\, United States
CATEGORIES:Graduate Seminar
END:VEVENT
END:VCALENDAR