BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
X-WR-CALNAME:Daniel Bienstock - Columbia University
X-WR-TIMEZONE:Eastern Time (US & Canada)
BEGIN:VEVENT
DTSTAMP:20260512T062221Z
UID:tag:localist.com\,2008:EventInstance_52259420272803
DTSTART:20260324T201500Z
DTEND:20260324T211500Z
DESCRIPTION:ORIE Colloquium and Data Science Distinguished Lecture\n\n \n\n
 Title: De-Risking Solutions to Optimization Problems\n\n \n\nDaniel Bienst
 ock \n\nBio: Daniel Bienstock is Liu Family Professor of Operations Resear
 ch at Columbia University\, with a joint appointment in Applied Math and a
  courtesy appointment in Electrical Engineering.  He received the PhD in O
 perations Research from MIT. His work focuses on methodology and computati
 onal aspects of optimization\, with additional focus on large-scale applic
 ations of optimization in engineering domains.\n\n Abstract:  Optimal or n
 ear-optimal solutions to real-world optimization problems tend to exhibit 
 what might be termed 'concentration' of actions or of deployed resources. 
  This is due to the nature of the task at hand (i.e.\, optimization) which
  conspires with flexibility present in real systems.  Thus\, as a simple e
 xample\, in an optimal solution to a  logistical problem we might see heav
 y usage of a certain combination of routes by large vehicles.  Such concen
 trations are a source of risk -- they represent a focused point of failure
  which the optimization problem was not tasked to protect from.\n\n In pri
 nciple\, one could try to contain concentration by imposing 'capacities' o
 n features at risk\; however\, there may well be too many such potential f
 eatures and it is not even clear\, a priori\, what the capacities should b
 e.\n\nIn this talk we present a strategy to address these issues that stem
 s\, in part\, from an attitude frequently found among real-world stakehold
 ers\; namely that the overt imposition of a risk structure in the optimiza
 tion task is\, often\, anathema -- it amounts to prioritizing conservatism
  over solution performance.\n\n Rather than attempting to solve\, directly
 \, a risk-aware problem (e.g.\, a stochastic or robust optimization proble
 m) we describe a rapid procedure that either (a) adjusts a given optimal s
 olution so as to substantially reduce its risk while\, also\, not substant
 ially increasing cost\, or (b) demonstrates that (a) is not possible.  Our
  procedure relies on first-order concepts and a cutting-plane algorithm.  
 We describe numerical experiments at scale.\n\n Joint work with Blake Siss
 on (Columbia) and Remi Akinwonmi and Alexandra Newman (Colorado School of 
 Mines).\n\n-------------------------------------\n\nThis talk is co-sponso
 red with the School of Operations Research and Information Engineering and
  the Center for Data Science for Enterprise and Society.
GEO:42.443451;-76.481506
LOCATION:Frank H. T. Rhodes Hall\, 253
SUMMARY:Daniel Bienstock - Columbia University
URL;VALUE=URI:https://events.cornell.edu/event/daniel-bienstock-columbia-un
 iversity
CATEGORIES:Colloquium
END:VEVENT
END:VCALENDAR