Graduate Student Seminar
Wednesday, February 26, 2014 · 11 AM - 12 PM
Speaker | Juyoung Jeong |
Session Chair | Zois Boukouvalas |
Discussant | Dr. Potra |
Place | BS 120 |
- Title
- Convex cones of matrices
- Abstract
- Convex cones of matrices—especially positive semidefinite (PSD) matrices, doubly nonnegative (DNN) matrices, or completely positive (CP) matrices— and its dual cone arise in several areas of applied mathematics, including optimization. We say a convex cone K is pointed if the intersection of K and –K is only the zero set, and K is solid if its interior is nonempty. If a convex cone K is closed, pointed, and solid, then K is said to be a full cone. We: (i) show that PSD, DNN, CP are all full cones; (ii) demonstrate their duals explicitly; and (iii) investigate the difference between DNN and CP.