Component
École Nationale Supérieure d'Électrotechnique d'Électronique d'Informatique d'Hydraulique et des Télécommunications
Objectives
Understand, know how to evaluate (complexity, efficiency, precision) and use the basic tools of numerical linear algebra.
Description
Singular value decomposition, pseudo-inverse of a matrix and applications.
Notions of numerical errors (direct and inverse errors) and conditioning of a matrix.
Dense matrix factorization for solving linear systems: LU, Cholesky, QR.
Iterative algorithms for solving linear systems: relaxation methods (Jacobi, Gauss-Seidel), steepest descent and conjugate gradient.
Algorithms for the search of eigenvalues/vectors : iterated power, Jacobi algorithm.