Séminaire en format hybride à HEC Montréal, salle Hélène-Desmarais ou Zoom.

Stochastic integer programming (SIP) problems combine the power of integer decision variables for modeling discrete decisions and logical relationships with the ability of stochastic programming to manage uncertainty when operating, planning, and designing systems. Because of this combination, SIP can be useful in a wide range of applications including power grid operation, employee staffing, and supply chain network design. This combination of features also leads to models that can be extremely difficult so solve. We present recent work in solving these types of problems, including the enhancements to the branch-and-cut method for solving a single instance and techniques for accelerating the solution of multiple instances, as is required when using the sample average approximation technique. This is based on work with Rui Chen and Harshit Kothari.

Biographie : Dr. Jim Luedtke est professeur titulaire au département d’ingénierie industrielle et système de l’Université de Wisconsin-Madison. Il a obtenu son doctorat de Georgia Institute of Technology en 2007. Spécialiste de l’optimisation nonlinéaire, stochastique et linéaire avec entiers mixtes, il est actif au sein de la société INFORMS d’optimisation et rédacteur adjoint de Mathematical Programming Computation.