[Seminario] INVITACIÓN COLOQUIO OPTIMIZACIÓN MIÉRCOLES 09 DE MARZO A CONTAR DE LAS 15:00 HRS.
Maria Ines
mrivera en dim.uchile.cl
Lun Mar 7 15:35:47 CLT 2016
*Estimados Académicos y Alumnos, *
*
Se les invita para este miércoles 09 de Marzo a contar de las 15:00
hrs, al COLOQUIO OPTIMIZACIÓN, TITULADO**"A mediodía de Optimización y
análisis variacional", el cual tendrá lugar en la de Seminarios John
Von Neumann del CMM. *
**
*EXPOSITORES*
- Pr. Dimitri Bertsekas, Departamento de Ingeniería Eléctrica e
Informática en la Facultad de Ingeniería en el Instituto de Tecnología
de Massachusetts (MIT), en Cambridge, Massachusetts.
- Pr. Assen Dontchev, Mathematical Reviews y la Universidad de Michigan.
- Pr. R. Tyrrell Rockafellar, Departamentos de matemáticas y matemáticas
aplicadas en la Universidad de Washington, Seattle.
*PROGRAMA
Miércoles, 9 de marzo:
*
*/- 15:00-15:45 hrs:/*
*//*
*/Pr. Dimitri Bertsekas, MIT, Massachusetts./*
*Title : Enhanced Fritz John Optimality Conditions and Sensitivity Analysis
Abstract:*We discuss a set of Fritz John conditions for convex and
nonconvexconstrained optimization, which are more powerful than the
classicalones, and assert the existence of multipliers with some
minimalsensitivity properties under very general assumptions. We use
theseconditions to construct a unifying constraint qualification
frameworkbased on the notion of pseudonormality. We also use them to
construct ahierarchy of Lagrange multipliers with increasingly stronger
sensitivity bproperties, culminating with results that assert the
existence ofmultipliers with optimal sensitivity properties, in the
sense that theyquantify the optimal rate of cost function improvement as
constraintsare violated.
*/
- 15:45-16:30 hrs./*
*/Pr. Assen Dontchev, Mathematical Reviews and University of Michigan.
Title: The Newton method: main developments and enhancements.
/*
*/Abstract:/*This talk gives an overview of convergence results for
various Newton-type methods including extensions to variational
inequalities and optimization problems based on metric regularity
properties of mappings in variational analysis. Applications to model
predictive control will be presented.
*/
16:30-17:00 hrs: Coffee break and discussion/*
*-17:00-17:45 hrs.*
*/Pr. R. Tyrrell Rockafellar, University of Washington, Seattle
Title: Solving Stochastic Variational Inequalities
/*
*/Abstract:/*With Monotonicity by Progressive Hedging Stochastic
variational inequality problems, single-stage or multistage, canbe
formulated with nonanticipativity as an explicit constraint on the
interaction of decisions with information. The introduction
ofmultipliers for that constraint allows in theory for a decomposition
into separate subproblems each of the scenarios, as if hindsight were
available, but a key question is how to generate the right value for
those multipliers.
When monotonicity is present in the variational inequality, the
progressive hedging algorithm of stochastic programming can be adapted
to this setting. Solutions can then be calculated by iterations in which
the multipliers, in parallel to those providing the present cost of
future information in simpler models, are approximated and adjusted step
by step.
Esperando contar con su presencia les saluda,
Ma. Inés Rivera
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listas.dim.uchile.cl/pipermail/seminario/attachments/20160307/466d8690/attachment-0001.html>
Más información sobre la lista de distribución Seminario