OPTIMIZATION SEMINAR



An Introduction to Positive Semidefinite Programming




Allen Holder
University of Colorado at Denver
Department of Mathematics

Tuesday, Oct. 28, 1997, 12:00 noon
CU-Denver Bldg., Room 626


ABSTRACT:

Positive semidefinite programming is a recent topic of interest for many researchers. Much of this research is supported by the fact that positive semidefinite programs are solvable by interior point methods. This talk discusses not only these solution techniques, but other topics such as duality and problem formulation.

After stating a few problem formulations, we motivate why PSD programs are of interest by presenting several applications. Convex conic programming is used to develop the needed duality theory. Finally, the framework for primal-dual interior point algorithms is presented.