6.337J Introduction to Numerical Methods
Advanced introduction to numerical linear algebra and related numerical methods. Topics include direct and iterative methods for linear systems, eigenvalue and QR/SVD factorizations, stability and accuracy, floating-point arithmetic, sparse matrices, preconditioning, and the memory considerations underlying modern linear algebra software. Starting from iterative methods for linear systems, explores more general techniques for local and global nonlinear optimization, including quasi-Newton methods, trust regions, branch-and-bound, and multistart algorithms. Also addresses Chebyshev approximation and FFTs. MATLAB is introduced for problem sets.
6.337J will not be offered this semester. It will be available in the Spring semester, and will be instructed by S. G. Johnson.
This class counts for a total of 12 credits. This is a graduate-level class.
You can find more information at the Optimization at MIT: 6.337J site.
© Copyright 2015 Yasyf Mohamedali