COSC 271 Numerical Linear Algebra (Formerly COSC 240)

The course examines in the context of modern computational practice algorithms for solving linear systems Ax = b and Az = λx. Matrix decomposition algorithms, matrix inversion, and eigenvector expansions are studied. Algorithms for special matrix classes are featured, including symmetric positive definite matrices, banded matrices, and sparse matrices. Error analysis and complexity analysis of the algorithms are covered. The algorithms are implemented for selected examples chosen from elimination methods (linear systems), least squares (filters), linear programming, incidence matrixes (networks and graphics), diagonalization (convolution), sparse matrices (partial differential equations).

Cross Listed Courses

ENGS 106 and MATH 116

Prerequisite

COSC 71, MATH 26, or ENGS 91. Students are to be familiar with approximation theory, error analysis, direct and iterative techniques for solving linear systems, and discretization of continuous problems to the level normally encountered in an undergraduate course in numerical analysis.