IE 614 Nonlinear
Programming Spring 2003
Instructor:
M. C. Pinar
Room: 305
Office Hours:
To be announced.
Textbook:
Topics to
be covered:
- Generalities
- Unconstrained Optimization:
Necessary and Sufficient Conditions for Local Minima, Role of Convexity
- Quadratic Unconstrained
Problems, Existence of Optimal Solutions, Iterative Computational Methods
- Gradient Methods,
Details and Convergence Issues
- Rate of Convergence
Issues
- Newton's Method and
Variants, Least Squares and Gauss-Newton Method
- Conjugate Direction
Methods, Quasi-Newton Methods, Non-derivative Mehtods
- Constrained Optimization:
Optimization over a Convex Set, Optimality Conditions
- Feasible Directions
Method, Gradient Projection Method
- Equality Constrained
Problems, Basic Lagrange Multiplier Theorem
- Equality Constrained
Problems, Sufficient Conditions, Augmented Lagrangians, Sensitivity Issues
- Inequality Constrained
Problems, Karush-Kuhn-Tucker Optimality Conditions, Fritz-John Conditions,
Constraint Qualifications
- Convex Cost and Linear
Constraints, Duality Theorem, LP and QP duality
- Conic Duality and Semidefinite
Programming
- Barrier and Interior
Point Methods
We will follow closely the textbook in the order of topics specified above. This is an introductory course in nonlinear
programming. Basic knowledge of linear algebra and some background on optimization
are assumed.
Grading Policy
Grading will be based on occasional
homework (50%) and a midterm test (25%) and a final examination (25%), the
dates of which will be announced.
All written submitted work in hws and exams are assumed to be strictly personal.
Any behavior in violation of the University Policy on Cheating is strongly
discouraged.