21-344 Numerical Linear Algebra

9 units, Carnegie Mellon University

An introduction to algorithms pertaining to matrices and large linear systems of equations. Direct methods for large sparse problems including graph data structures, maximum matchings, row and column orderings, and pivoting strategies. Iterative methods including Conjugate Gradient and GMRES, with a discussion of preconditioning strategies. Additional topics include: computation of eigenvalues and eigenvectors, condition numbers, the QR and singular value decompositions, least-squares systems. (Three 50 minute lectures)

Prerequisites: (15-112 or 02-120) and (21-241 or 21-240 or 21-242) and (21-259 or 21-268 or 21-269)

Offered: Spring 2026

Courseweave needs JavaScript to show sections, workloads and planning tools.