CENG 504

Optimization Methods

Linear programming, nonlinear programming, iterative methods and dynamic programming are presented, especially as they relate to optimal control problems. Discrete and continuous optimal regulators are derived from dynamic programming approach which also leads to the Hamilton-Jakobi-Bellman Equation and the Minimum Principle. Linear quadratic regulators, linear tracking problems and output regulators are treated. Linear observer and the separation theorem are developed for controller implementation.

Topics
Introduction to optimization
Mathematical Review I : Vectors and Matrices
Mathematical Review II : Calculus
Unconstrained optimization, tek boyutlu arama
golden section, Fibonacci, Newton s method
gradient search methods: Steepest-descent, Newton s method, Least squares analysis
Constrained optimization
Linear Programming
Evaluation and Review
Non-derivative optimization, heuristic optimizastion methods
Neural networks, simulated annealing
Term project presentations and discussions I
Term project presentations and discussions II
Term project presentations and discussions III