CS51500 - Numerical Linear Algebra

Fall 2015

Days/Time: TTh / 10:30-11:45 am
Credit Hours: 3

Learning Objective:
None

Description:
Direct and iterative solvers of dense and sparse linear systems of equations, numerical schemes for handling symmetric algebraic eigenvalue problems, and the singular-value decomposition and its applications in linear least squares problems.

Topics Covered:
Dense Matrix Computation: Direct linear system solvers; LU and Cholesky factorization schemes, Norms and condition numbers, Pivoting strategies, scaling, and iterative refinement. Least squares problems; Orthogonal projections, Orthogonal factorization schemes--Givens, Householder, and Gram-Schmidt, Singular-value decomposition. The symmetric eigenvalue problem; Eigenvalues and eigenvectors, Power method and inverse iteration, Reduction to the tridiagonal form, Extraction of eigenpairs. Iterative methods for sparse linear systems: Discretization of partial differential equations, Sparse matrices, Basic iterative linear system solvers, Projection methods, Krylov subspace methods, Schemes for normal equations, Preconditioning techniques.

Prerequisites:
A bachelor degree in computer science or an equivalent field. Students not in the Computer Science master's program should seek department permission to register.


Homework:
None

Textbooks:
Official textbook information is now listed in the Schedule of Classes. NOTE: Textbook information is subject to be changed at any time at the discretion of the faculty member. If you have questions or concerns please contact the academic department.


Computer Requirements:
None

ProEd Minimum Requirements: view

Tuition & Fees: view

INSTRUCTOR

David Gleich
Phone
765-496-7300
Email
dgleich@purdue.edu
Office
Purdue University
Department of Computer Science
305 N. University Street
West Lafayette, IN 47907
Instructor HomePage

You May also be Interested In: