George Mason University
CSI/Statistics Colloquium Series
Seminar Announcement


Nonlinear Rescaling in Constrained Optimization

Roman A. Polyak


George Mason University


ABSTRACT
The nonlinear rescaling principle (NRP) consists of transforming the objective function and/or the constraints of a given constrained optimization problem into an equivalent problem and using the classical Lagrangian of the equivalent problem for both theoretical analysis and numerical methods. The nonlinear transformation is parameterized by a positive scalar parameter or by a vector of scaling parameters one for each constraint. The nonlinear rescaling (NR) methods consists of sequential unconstrained minimization of the classical Lagrangian for the equivalent problem, followed by an explicit formula for the Lagrange multipliers update.

We will focus on the primal-dual aspects of NR methods. Convergence, rate of convergence, complexity issues as well as applications and numerical results will be discussed.


Friday, September 10, 1999
George W. Johnson Center, Assembly Room C
Seminar at 10:45 a.m.
Refreshments at 10:30 a.m.
For the 1999 Fall Seminar Schedule, go to
http:www.science.gmu.edu/statseminars