Optimisation methods

Organisation

Cours/TD/TP/projet/examen : 45

Acquis d'apprentissage

Acquiring some notions of optimization in continuous, discrete or mixed spaces and their relationship with concrete applications.

Prérequis

Basic Calculus, Basic Algebra

Contenu

- Dynamic programming
- Branch and Bound methods
- B&B and the Travelling Salesman problem : the Little algorithm
- Linear Programming : the simplex algorithm
- Unconstrained non-linear Programming : gradient methods, Newton method, quasi-Newton methods
- Metaheuristics for hard optimization : Taboo Search, Evolutionary Computation, Simulated Annealing
- Applications to Pattern Recognition : elastic distance, Dynamic Time Warping, gradient methods in neural networks, etc.

Evaluation

continuous exam
written examination

Bibliographie

- J.Dréo, A.Petrowski, P.Siarry, E.Taillard, Metaheuristics for hard optimization, Mathematical Methods for Operations Research, vol. 66, N°3, December 2007, Springer Berlin.
- L.Rabiner, B.H. Juang, Fundamentals of Speech Recognition, Prentice Hall Signal Processing Series, 1993.
- S.Haykin, Neural Networks : a comprehensive founda j