[NOVA Math] Seminar of Operations Research

The Center of Mathematics and Applications (NOVA Math), promote the Seminar of Operations Research with the title: "Eigenvector-based acceleration strategies for gradient-type methods". Marcos Raydan (Center for Mathematics and Applications (NOVA Math), NOVA FCT) is the speaker.

Abstract:
Several strategies are described and analyzed to speed-up gradient-type methods when applied to the minimization of strictly convex quadratics and strictly convex functions. The proposed techniques focus on relaxing the traditional optimal step length associated with gradient methods, including the steepest descent (SD) and the minimal residual (MR) methods. Such a relaxation avoids the well-known negative zigzag effect and allows the
iterates to move in the entire space which in turn implies that every so often the search direction approaches some eigenvector of the underlying Hessian matrix. The proposed speedups then rely on taking advantage of the properties of the Lanczos method once a search direction that approaches an eigenvector has been identified in order to accelerate the convergence towards the global minimizer. After analyzing the proposed strategies, we illustrate them on the global minimization of strictly convex functions.

May 20 (Wednesday), 15h00, Room 204 - Building II.