School Seminars and Colloquia

Extrapolation algorithm for convex feasibility problems with application to Semidefinite Programming

ORSUM Seminar

by Associate Professor Serge Kruk


Institution: Oakland University, Michigan
Date: Mon 26th March 2007
Time: 1:00 PM
Location: Room 213, Richard Berry Building, The University of Melbourne

Abstract: The convex feasibility problem under consideration is to find a common point of a countable family of closed affine subspaces and convex sets in a Hilbert space. We describe a general parallel block-iterative algorithmic framework in which the affine
subspaces are exploited to introduce extrapolated over-relaxations.
This framework encompasses a wide range of projection, subgradient projection, proximal, and fixed point methods encountered in various
branches of optimization. Numerical experiments in the context of large scale semidefinite programming are provided to illustrate the
benefits of the extrapolations.

For More Information: Mark Fackrell: mfackrel@ms.unimelb.edu.au