You are on page 1of 3

NUMERICAL ANALYSIS SUMMARY

Background :
Horner’s method; computer arithmetic and roundoff
errors; Taylor series.
Solution of Nonlinear Equations f (x) = 0
• Bisection Method: algorithm; convergence analysis;
guarenteed linear convergence en = O(2−n).
• Fixed Point Method xn+1 = g(xn);
convergence analysis; linear convergence en = O(|g 0(r)|n).
• Newton Method: xn+1 = xn − f (xn)/f 0(xn);
quadratic convergence en ≈ Ke2n
−xn−1
• Secant Method: xn+1 = xn − f (xn) f (xxnn)−f (xn−1 ) ,
superlinear convergence with en+1 ≈ Ke1.62n
• Forward and Backward errors and sensitivity.
Numerical Solution of Ax = b
• Gaussian Elimination (GE), and
computational complexity (# flops).
• LU factorization solution of Ax = b
• Condition Number for Ax = b, error bound.
• Swamping, and pivoting for GE and LU.
• Cholesky Factorization method for
symmetric positive definite systems.
• Iterative Methods: Jacobi, Gauss-Seidel, SOR:
error analysis, convergence, efficiency.
NUMERICAL ANALYSIS SUMMARY
Interpolation
n n x − xj
• Lagrange Formula: Pn(x) = f (xi) .
X Y

i=0 j=0,6=i xi − xj
• Divided Differences and Divided Difference Formula
n i
Pn(x) = f [x0, x1, . . . , xi] (x − xj ).
X Y

i=0 j=0
Construction of divided difference tables.
nf (n+1)(ξ)
• Interpolation Error: f (x)−Pn(x) = (x−xi) .
Y

i=0 (n + 1)!
• Runge Phenomenon
• Chebyshev Interpolation and error bounds
• Cubic Splines: properties and purpose
Least Squares
• Overdetermined Systems and the Normal Equations
• Least Squares and Data Fitting
• Ill-Conditioning and Least Squares
• QR Factorization solutions
Numerical Differentiation and Integration
• Differentiation Methods: basic formulas,
Taylor series error analysis, instability.
• Extrapolation: basic formulas for even-power
and all-power error series.
• Trapezoidal and Simpson Rules: composite formulas,
errors, degrees of precision.
• Romberg Integration from extrapolation.
• Gaussian Rules: transformation, degree of precision.
2
NUMERICAL ANALYSIS SUMMARY
ODE IVP Numerical Solutions :
for y 0 = f (t, y), y(a) = y0, t ∈ [a, b].
• Euler’s Method: wi+1 = wi + hf (ti, wi);
error, convergence, stabilty
• Taylor Methods: formulas; truncation errors, order
• Runge-Kutta Methods: general form, how to use;
explicit Trapezoidal method
wi+1 = wi + h(f (ti, wi) + f (ti + h, wi + hf (ti, wi))/2.
• Methods for ODE IVP Systems:
how to use basic explicit methods;
nth order ODE → to 1st order system.
• Stiff equations and implicit methods:
backward Euler and implicit Trapezoidal methods.

You might also like