BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VTIMEZONE
TZID:Europe/Vienna
BEGIN:DAYLIGHT
DTSTART:20190331T030000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20191027T020000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260404T081118Z
UID:5ce7b08ac6d7b622245155@ist.ac.at
DTSTART:20190527T140000
DTEND:20190527T150000
DESCRIPTION:Speaker: Martin Krejca\nhosted by Dan Alistarh\nAbstract: Estim
 ation-of-distribution algorithms (EDAs) are randomized search heuristics (
 RSHs) with successful applications in the domain of combinatorial optimiza
 tion. In contrast to other RSHs\, EDAs utilize a probabilistic model of th
 e problem domain\, which they evolve iteratively in the following way. In 
 each iteration\, the model is used in order to generate random solutions. 
 These samples are filtered and then used in order to refine the model such
  that it better incorporates the filtered samples.In this talk\, we focus 
 on univariate EDAs\, that is\, algorithms with a probabilistic model that 
 assumes independence among the problem variables\, and how they are analyz
 ed theoretically. To this end\, we introduce the general framework for suc
 h EDAs as well as common benchmark functions considered in the theory comm
 unity. Further\, we discuss some of our results that are concerned with th
 e expected run time of EDAs. Especially\, we learn about a drawback that m
 any univariate EDAs exhibit and that hampers the optimization process\, an
 d we see how this problem can be overcome.
LOCATION:Mondi Seminar Room 1\, Central Building\, ISTA
ORGANIZER:lmarr@ist.ac.at
SUMMARY:Martin Krejca: Theoretical Analyses of Estimation-of-Distribution A
 lgorithms
URL:https://talks-calendar.ista.ac.at/events/1992
END:VEVENT
END:VCALENDAR