
MODEL IDENTIFICATION, CALIBRATION AND DATA ANALYSIS
Prerequisites:

None 
Course unit contents:

1. Review of linear algebra concepts;
2. Iterative methods for the solution of large, sparse linear systems: a) conjugate gradient methods for symmetric systems; b) projection methods for nonsymmetric systems (GMRESBiCGSTAB); c) preconditioning; incomplete factorizations; sparse factorized approximate inverses; d) implementation techniques; sparse (CSR) matrix storage;
3. Methods for the calculation of eigenvalues and eigenvectors: a) Power and inverse power (with shift) methods; b) QR method.
4. Newton methods for nonlinear systems: a) derivation of the Newton methods; b) local convergence properties and introduction to globalization techniques; c) Picard method; d) implementation of the NewtonKrylov and inexact Newton methods.
5. The calibration as an ill posed problem;
6. Penalizing functions;
7. Likelihood method for estimation;
8. Generalized Method of Moments;
9. Deterministic and stochastic algorithms. 

