Tuesday, July 15

(This session will run until 5:45PM)

3:15 PM-5:45 PM
Meyer Library, Forum Room

New Methods for Least Squares Problems with Uncertainty and Structure

The celebrated least squares problem has undergone many developments recently, such as constrained least squares, total least squares, structured/constrained total least squares, etc. These developments share the common idea of taking into account some a priori knowledge in the data A,b, to define a "solution" to an overdetermined system Ax = b.

The purpose of the minisymposium is to present new methods that aim at taking into account different kinds of a priori information on the data and/or the solution. Problems with unknown-but-bounded (structured) uncertainty on the data, structured condition numbers, or recent techniques for large-scale problems are among these new developments. These problems are often linked with robustness and/or regularization issues.

Organizers: Laurent El Ghaoui and Herve Lebret
Ecole Nationale Superieure de Techniques Avancees, France

3:15 Parameter Estimation with Bounded Data Uncertainties
Ali H. Sayed, University of California, Los Angeles; Gene H. Golub and S. Chandrasekaran, Stanford University
3:45 Tikhonov Regularization for Large Scale Problems
U. Von Matt, ETH-Zentrum, Switzerland
4:15 Second-Order Cone Programming
L. Vandenberghe, Miguel Sousa Lobo, and Stephen Boyd, Stanford University; and Herve Lebret, Organizer
4:45 Robust Solutions to Least-Squares Problems with Uncertain Data
Laurent El Ghaoui and Herve Lebret, Organizers
5:15 Stability of Householder QR Factorization for Weighted Least Squares Problems
Anthony J. Cox and Nicholas J. Higham, University of Manchester, United Kingdom

AN97 Homepage | Program Updates|
Registration | Hotel and Dormitory Information | Transportation | Program-at-a-Glance | Program Overview

TMP, 4/4/97
tjf, 5/28/97
MMD, 5/30/97