Computer Science 510 - Numerical Analysis
FALL 2011
Course web page: http://www.cs.rutgers.edu/~richter/cs510
Instructor:
Gerard Richter
richter@cs.rutgers.edu, 732-445-2001 x2097
314 Core Building
office hrs: Monday 12-1 pm, Tuesday 1:30-3 pm, or by appointment
Teaching assistant:
Runhe(Bruce) Tian
rtian@cs.rutgers.edu
Hill 402
office hours: Monday, 3-5 pm, or by appointment
Texts (reference) - first two on reserve in Math Sciences Library
(Hill Center):
-
M. T. Heath, Scientific Computing, An Introductory Survey,
2nd edition, McGraw-Hill, 2002 (primary reference -
available in bookstore)
- G. Dahlquist & A. Bjorck, Numerical Methods,
Prientice-Hall, 1974, 2003,
2008
- G. H. Golub & C. F. Van Loan, Matrix Computations, 3rd
edition, Johns Hopkins University Press, Baltimore, 1996.
-
Matlab tutorial + links to other references
Prerequisites: calculus, linear algebra, ability to
program in a high level language, e.g., Matlab, Fortran, C
Grading:
-
written homework, computer programs ~ 20%
-
midterm ~ 35%
-
final exam ~ 45%
Objectives: derivation, analysis, implementation of
algorithms for numerical problems
Outline of topics...
- Floating point numbers and roundoff error
- Solution of nonlinear algebraic equations
- bisection method, regula falsi, fixed point iteration,
secant method, Newton's method
- convergence rates (linear, quadratic)
- systems of nonlinear equations - Newton's method
- Solution of linear algebraic systems
- Gaussian elimination/ LU decomposition
- special cases: symmetric, banded, sparse matrices
- error analysis, norms, condition number
- iterative methods (SOR, convergence rates)
- algorithms for parallel computers
- overdetermined systems, least squares solutions
- Other numerical linear algebra topics + applications
- QR decomposition
- Singular value decomposition
- Web search - PageRank
- Interpolation, approximation of functions
- the interpolating polynomial (its construction and error
term)
- piecewise polynomial interpolation, splines
- Numerical differentiation and integration
- quadrature formulas, error terms
- adaptive quadrature, Gaussian quadrature
- numerical differentiation, error terms
- Numerical solution of ordinary differential equations
- basic methods (Taylor methods, Runge-Kutta methods,
multistep methods)
- stability, consistency, convergence
- higher order equations, systems