BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Seminar by Prof. Angelos Georghiou\, McGill University
DTSTART:20190703T150000
DTEND:20190703T163000
DTSTAMP:20260916T004930Z
UID:5a14e901b1c857e5f9ed8662e886f2b6f0d8918e8c828c94c6a9decb
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof. Angelos Georghiou\, McGill University\nRobust Optimizati
 on with Decision-Dependent Information Discovery\n\nAbstract\n\nRobust opt
 imization is a popular paradigm for modeling and solving sequential decisi
 on-making problems affected by uncertainty. Most approaches assume that th
 e uncertain parameters can be observed for free and that the sequence in w
 hich they  are revealed is independent of the decision-maker’s actions.
  Yet\, these assumptions fail to hold in many applications where the time 
 of information discovery is decision-dependent and uncertain parameters on
 ly become observable after a costly investment.\nIn this paper\, we consid
 er two-stage robust optimization problems in which (part of) the decision 
 variables control the time of information discovery. Thus\, information av
 ailable at any given time is decision-dependent and can be discovered by m
 aking strategic exploratory investments in previous stages. We propose a n
 ovel dynamic formulation of the problem. We prove correctness of this form
 ulation and leverage our new model to provide a solution method inspired f
 rom the K-adaptability approximation approach\, whereby K candidate strate
 gies for each decision stage are chosen here-and-now and\, at the beginnin
 g of each period\, the best of these strategies is selected after the port
 ion of the uncertain parameters that was chosen to be observed is revealed
 . We reformulate the problem as an MILP solvable with off-the-shelf solver
 s\nand demonstrate its effectiveness on both synthetic and real data insta
 nces of the active preference elicitation problem used to learn the moral 
 priorities of policy-makers in charge of allocating housing resources to t
 he homeless.\n\n 
LOCATION:ODY 0 03 https://plan.epfl.ch/?room=ODY003
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR