Campuses:

Branch-and-cut for cardinality constrained optimization

Monday, July 25, 2005 - 2:30pm - 3:15pm
EE/CS 3-180
Ismael de Farias JR. (University at Buffalo (SUNY))
Given a set of continuous variables and for each variable a constant,
a cardinality constraint requires that no less than a specified number
of variables is equal to their corresponding constants. Cardinality
constraints appear in several applications, such as feature selection
in data mining. I will present a polyhedral study of problems with
cardinality constraints and a branch-and-cut approach to solve them.
In particular, I will present a few unusual properties of these
polyhedra.

Joint work with Ming Zhao.