instituto de matemáticas universidad de sevilla
Antonio de Castro Brzezicki
Nonsmooth optimization: conditioning, convergence, and semi-algebrais models
Seminario Doctorado
PhD Program Activity
Date: 19.10.2016 From 10.00 till 12.00
Venue: Seminario I (IMUS), Edificio Celestino Mutis
Variational analysis has come of age.  Long an elegant theoretical toolkit for variational mathematics and nonsmooth optimization, it now increasingly underpins the study of algorithms, and a rich interplay with semi-algebraic geometry illuminates its generic applicability.  As an example, alternating projections - a rudimentary but enduring algorithm for exploring the intersection of two arbitrary closed sets - concisely illustrates several far-reaching and interdependent variational ideas.  A transversality measure, intuitively an angle and generically nonzero, controls several key properties: the method's linear convergence rate, a posteriori error bounds, sensitivity to data perturbations, and robustness relative to problem description.  These linked ideas emerge in a wide variety of computational problems.  Optimization in particular is rich in examples that depend, around critical points, on “active" manifolds of nearby approximately critical points.  Such manifolds, central to classical theoretical and computational optimization, exist generically in the semi-algebraic case.  We discuss examples from eigenvalue optimization and stable polynomials in control systems, and a prox-linear algorithm for large-scale composite optimization applications such as machine learning.
(This lecture is a version of an invited talk at the 2014 International Congress of Mathematicians in Seoul.)

Dirigido especialmente a alumnos que estén trabajando en las áreas de Optimización, Análisis No Lineal o Lógica

Las inscripciones para esta actividad se realizarán por correo electrónico a acti2-imus@us.esindicando lo siguiente:

Nombre de la actividad:

Nombre completo del participante:

DNI  o pasaporte:

Correo electrónico: