Continuation in optimization: From interior point methods to Big Data
Wednesday 1 November, 12:10 in M/0.34 (MATHS Building)
In this talk, Professor Jacek Gondzio will discuss similarities between two homotopy-based approaches:
- (inexact) primal-dual interior point method for LP/QP, and
- preconditioned Newton conjugate gradient method for big data optimisation.
Both approaches rely on clever exploitation of the curvature of optimised functions and deliver efficient techniques for solving optimisation problems of unprecedented sizes. We will address both theoretical and practical aspects of these methods applied to solve various inverse problems arising in signal processing.
Part of this work was done jointly with former PhD student, Kimonas Fountoulakis.