BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:IC Mondays seminars - Approximation in Multiobjective Optimization
  
DTSTART:20100412T161500
DTSTAMP:20260408T070953Z
UID:0ecc441befd3a45850e1db975d921c3d1bb079155cabb052d8031d7f
CATEGORIES:Conferences - Seminars
DESCRIPTION:Dr.Ilias Diakonikolas\, Columbia University\, Computer Science
  Department\nAbstract\nIn multiobjective optimization\, solutions to an op
 timization problem are evaluated with respect to several cost criteria\, a
 nd we are interested in a class of solutions that capture the trade-off be
 tween these objectives\, the so-called Pareto curve. The problem is that t
 he Pareto curve is typically prohibitively large (or even infinite) and he
 nce we cannot construct the full curve. Instead\, we want to efficiently c
 ompute a small set of solutions (as small as possible) that provides a "go
 od enough" representation (as good as possible) of the whole design space.
  This has been the underlying goal of much of the research in the area\, w
 here many heuristics have been proposed for this purpose\, usually however
  without any performance guarantees or complexity analysis.\n\nIn recent y
 ears\, we initiated a systematic investigation to develop the theory of mu
 ltiobjective approximation along similar rigorous lines as the approximati
 on of single-objective problems. We address the problem of efficiently com
 puting a succinct approximation to the Pareto curve using as few points as
  possible. If we are to select only a certain number of solutions\, how sh
 all we pick them so that they represent as accurately as possible the whol
 e spectrum of possibilities?\n\nI will present general methods for computi
 ng efficiently succinct approximations to the Pareto curve for broad class
 es of problems\, and show how they can be applied to particular well-studi
 ed problems (such as multiobjective shortest paths\, linear programming\, 
 etc)\n\n(The talk is based on joint works with Mihalis Yannakakis.)\n\nBio
 graphy\nIlias Diakonikolas received his undergraduate degree in Electrical
  and Computer Engineering from the National Technical University of Athens
 . He is currently pursuing a Ph.D. in Computer Science at Columbia Univers
 ity under the supervision of Professor Mihalis Yannakakis.\n\nIlias' resea
 rch interests include areas of theoretical computer science such as approx
 imation algorithms\, computational complexity\, computational learning and
  property testing. His thesis work focuses in the development of a coheren
 t algorithmic theory for multiobjective optimization problems. His work in
  this area has been honored by the INFORMS society.
LOCATION:INM 202
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR