COURSE OUTLINE:

Main topics covered :

1. General introduction to nonlinear programming methods    including the following algorithms for unconstrained optimization:

 - steepest descent,

 - conjugate gradient,

 - Newton, and pseudo Newton algorithms and solution of

   nonlinear vector equations.

2.  Introduction to the linear least squared error problem;

    projections and null spaces.

 

3. Introduction to constrained optimality :

 - development of the necessary condition for the convex case ,

 - convex duality; the perturbation function and dual function for

  equality constraints ,

 - extension to the case  of mixed equality and inequality constraints ,

 - feasible direction algorithm . 

4. Topics chosen  and presented by the students.

5. Introduction to nonlinear dynamic optimization :

 - existence theory of optimal control solutions ,

 - the relaxed control problem,

 - special properties of linear control problems,

 - the Pontryagin Maximum Principle,

 - problems in the calculus of variations,

-         sufficiency of the Maximum principle.

 

 

Texts :

Lecturer's notes are a basis for the course and will be

distributed in the class.

 

 

Reference texts:

1.  D. P. Bertsekas, "Nonlinear Programming", Athena Scientific Publisher,1995. [T57.8 B47 1995]

2.  D. P. Bertsekas, "Dynamic Programming and Optimal Control", Athena Scientific Publisher, 1995. [T57.83 B476 1995]

3.  E. Polak, "Computational Methods in Optimization", lecture notes handed out at University of California at Berkeley.

[QA402.5 P58]