Notes and Resources...

are linked from the schedule.

Catalog description

Numerical Optimization, Newton's Methods for nonlinear equations and unconstrained optimization. Global methods, nonlinear least squares, integral equations.

• 1. Math 524 (Linear Algebra)
• 2. Math 542 (Numerical Solution of Differential Equations), or
  
• 2. Math 543 (Numerical Matrix Analysis)

Computational Resources:

You must have access to a somewhat modern version of Matlab, or some other computational environment that you are comfortable using. Class accounts for the labs will be available. You can also use SDSU's Rohan Sun Enterprise system or another capable system. [How to open a ROHAN account].

Required Text and Reading Materials:

Numerical Optimization, Jorge Nocedal and Stephen J. Wright, Springer-Verlag, Springer Series in Operations Research, 1999. ISBN 0-387-98793-2.

Errata: [1st Edition, 1st Printing], [1st Edition, 2nd Printing], [2nd Edition]

Class web page [http://terminus.sdsu.edu/SDSU/Math693a_f2007/], and handouts.

Course Outline (as of 8/27/2007):

Nocedal/Wright: 1-Introduction, 2-Fundamentals of Unconstrained Optimization, 3-Line Search Methods, 4-Trust-Region Methods, 5-Conjugate Gradient Methods, 6-Practical Newton Methods, 8-Quasi-Newton Methods, 10-Nonlinear Least-Squares Problems, 11-Nonlinear Equations.
[To Be Revised]

Professor

Peter Blomgren
blomgren AT mail DOT SDSU DOT EDU, (619)594-2602
Class hours: TuTh 4:00p – 5:15p, BA-260
Office hours: TuTh 2:00p – 3:30p (GMCS-587), and by appointment.