BEGIN:VCALENDAR
VERSION:2.0
PRODID:researchseminars.org
CALSCALE:GREGORIAN
X-WR-CALNAME:researchseminars.org
BEGIN:VEVENT
SUMMARY:Michael Friedlander (*Friday*) (UBC Vancouver)
DTSTART:20261113T223000Z
DTEND:20261113T233000Z
DTSTAMP:20261117T011932Z
UID:SFUOR/79
DESCRIPTION:Title: <a href="https://researchseminars.org/talk/SFUOR/79/">W
 hen can you trust an optimal assignment?</a>\nby Michael Friedlander (*Fri
 day*) (UBC Vancouver) as part of PIMS-CORDS SFU Operations Research Semina
 r\n\nLecture held in K9509.\n\nAbstract\nThe assignment problem pairs work
 ers with tasks at the lowest total cost. But costs are often measured\, es
 timated\, and rounded. How much error can we tolerate? Classical sensitivi
 ty analysis furnishes an answer\, but only after the fact. We instead seek
  a bound a priori. Counterintuitively\, randomness helps: superpose a Gaus
 sian perturbation and\, whatever the costs\, the tolerable error admits a 
 lower bound that depends exclusively on the problem size\, the scale of th
 e perturbation\, and a preselected confidence level. This bound informs th
 ree choices: the strength of quadratic regularization\, the tolerance for 
 the auction algorithm\, and the requisite number of bits to store the data
 .\n
LOCATION:https://researchseminars.org/talk/SFUOR/79/
END:VEVENT
END:VCALENDAR
