Retour aux activités
Séminaire pas ordinaire

On integrating decision diagrams and mixed integer programming for optimization

iCalendar

24 oct. 2017   12h00 — 13h00

Jaime E. Gonzalez Polytechnique Montréal, Canada

In recent years, optimization based on decision diagrams have come up as an alternative technology that operations researchers should have in their toolbox. Decision diagrams (DDs) are graphical data structures which can compactly represent the solution set of a combinatorial problem. In this talk, we mainly discuss the basics of decision diagrams theory for optimization and then, a framework which integrates approximate DDs and mixed integer programming (MIP) technology for modeling and tackling optimization problems. In the framework, an approximate DD plays the role of a search tree by identifying parts of the search space that can be efficiently explored with MIP technology while dual bounds can also be obtained. Conversely, the MIP results are iteratively employed to refine the DD and find primal bounds.


Ce séminaire s'adresse seulement aux étudiants du GERAD.
Nous vous remercions de confirmer votre présence en indiquant votre nom complet. Des pizzas et des breuvages seront servis aux participants ou vous pouvez apporter votre lunch.

Luciano Costa responsable

Lieu

Salle 4488
Pavillon André-Aisenstadt
Campus de l'Université de Montréal
2920, chemin de la Tour Montréal QC H3T 1J4 Canada