4 présentations

• 
  10h30 - 10h55

  Dynamic emergency medical services network design: A novel probabilistic envelope constrained stochastic model

  • Peng Chun, prés., HEC Montreal
  • Erick Delage, GERAD, HEC Montréal
  • Jinlin Li, Beijing Institute of Technology

  This talk introduces a two-stage stochastic programming model for dynamic emergency medical services network design. This model enforces a minimum probabilistic coverage (either through chance constraints or probabilistic envelope constraints) of future emergency demand while minimizing total expected cost over a planning horizon. Numerical experiments involve data from Northern Ireland.

• 
  10h55 - 11h20

  Improving stroke routing protocol

  • Amir Ardestani-Jaafari, prés., McGill University
  • Beste Kucukyazici, McGill University

  Stroke is medical emergency and must be treated immediately. However, transporting patients to the closest stroke hospital may not be the best solution. This often causes congestion in some hospitals, while underutilization in others. We study patients routing protocol under congestion using robust queuing technique.

• 
  11h20 - 11h45

  Robust binary optimization with an application to talent analytics

  • Aurélie Thiele, prés., Southern Methodist University
  • Sean Barnes, University of Maryland College Park
  • Margret Bjarnadottir, University of Maryland College Park

  We consider a binary linear programming problem in the presence of high uncertainty on the objective coefficients. We investigate theoretically four models based on an "estimate-then-optimize" paradigm and a robust optimization approach. We finally compare the results of those four methods using real-life data about baseball team player selection.

• 
  11h45 - 12h10

  Preference robust optimization for decision making under uncertainty

  • Erick Delage, prés., GERAD, HEC Montréal
  • Jonathan Li, Telfer School of Management, University of Ottawa

  While different risk measures can account for risk aversion, it is often unclear which one models best a decision maker’s perception of risk. We introduce preference robust optimization as a way of accounting for ambiguity about the DM’s preferences. We illustrate numerically our findings with a portfolio allocation problem and discuss possible extensions.