BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:On relaxation Methods for Mathematical Programs with Complementari
ty Constraints
DTSTART;VALUE=DATE-TIME:20180709T030000Z
DTEND;VALUE=DATE-TIME:20180709T040000Z
DTSTAMP;VALUE=DATE-TIME:20200410T001742Z
UID:indico-contribution-1752@indico.math.cnrs.fr
DESCRIPTION:Speakers: Mounir HADDOU (Centre de mathématiques Institut Nat
ional des Sciences Appliquées de Rennes\, France.)\nWe propose a new fami
ly of relaxation schemes for mathematical programs with complementarity\nc
onstraints that extends the relaxations converging to an M-stationary poi
nt.\nWe discuss the properties of the sequence of relaxed non-linear progr
ams as well as stationarity\nproperties of limiting points. We prove under
a new and weak constraint qualification\, that\nour relaxation schemes ha
ve the desired property of converging to an M-stationary point.\nUnfortuna
tely\, in practice\, relaxed problems are only solved up to approximate st
ationary\npoints and the guarantee of convergence to an M-stationary point
is lost. \nWe define a new strong approximate stationarity condition and
prove that we can maintain\nour guarantee of convergence and attain the de
sired goal of computing an M-stationary point.\nA comprehensive numerical
comparison between existing relaxations methods is performed\nand shows pr
omising results for our new methods.\nWe also propose di↵erent extension
s to tackle MPVC ( vanishing constraints) and MOCC\n(cardinality constrain
ts) problems.\n\nhttps://indico.math.cnrs.fr/event/3023/contributions/1752
/
LOCATION:Ho Chi Minh City University of Science
URL:https://indico.math.cnrs.fr/event/3023/contributions/1752/
END:VEVENT
END:VCALENDAR