Nonlinear Optimization (M12)

Winter Term 2010/11

Lecturer: C. Helmberg

Wed 7:30 - 9:05, Room 2/NK003 (neu! statt Dienstag)

(includes exercises)

Content:

We discuss basic algorithmic approaches for solving smooth nonlinear optimization problems. Aspects of interest are convergence rate, computational efficiency and numerical behavior.
Unconstrained Optimization: Newton and quasi-Newton methods, line search, trust regions, conjugate gradients, approximate and automatic differentation
Constrained Optimization: Lagrange multipliers, quadratic programming, penalty, barrier and augmented Lagrangian methods, sequential quadratic programming.

Audience:

any Mathematics program starting from the 7th term (5th should be ok, too)

Requirements:

Linear Algebra, Analysis, Basics of Optimization