M.S. Comprehensive Examination Topics
Numerical Analysis
2006
Taylor Series (convergence proofs, truncated series)
Error Characterization (relative, absolute, big ``O'')
Computer Arithmetic
Floating point representations
Machine epsilon, roundoff error, loss of significance
Root Finding f:R --> R
Convergence theorems, modes of failure
Linear, quadratic, superlinear convergence
A priori convergence theorems
Solution of Linear Systems
System Theory
Matrix norms and condition number of a matrix
Ill-conditioned systems, real symmetric matrices
Special systems
Upper/lower triangular (diagonal dominance)
Tridiagonal and banded systems
Symmetric positive definite
Solution methods
Gaussian elimination: scaling, scaled partial pivoting
Backward/Forward substitution
Computational complexity: O(np)
Power method and convergence analysis
General Least Squares
Singular value decomposition: pseudoinverse and related theory
Interpolation
Existence and uniqueness theorems
Error analysis for even grids
Splines
Linear, quadratic, cubic, degree
Error analysis for natural cubic splines
Integration
Trapezoid, Simpson's rules
n-point Gaussian quadrature: basic rules, error analysis
Orthogonal polynomial derivation of Gauss n-point rules
Initial Value Problems: Existence, uniqueness, blowup
Initial Value Problem Methods: y'(t) = f(y(t),t)
Adams-Bashforth methods (explicit)
Adams-Moulton methods (implicit)
Runge-Kutta (RK) methods (explicit)
Backward difference methods
Euler, midpoint, implicit trapezoid, backward Euler
Initial Value Problem Methods: Theory
Absolute stability, A-stable
Second-Order Boundary Value Problems
Uniqueness and nonuniqueness of solutions
Shooting methods: Newton/secant
Finite difference methods
References:
Numerical Analysis, Richard L. Burden and J. Douglas Faires,
PWS-Kent Publishing Company.
Numerical Analysis, David Kincaid and Ward Cheney, Brooks/Cole.
©Davis, Vogel