Вы находитесь на странице: 1из 18

Introduction to Numerical Analysis II:

Course Introduction

Xiaozhe Hu

The Pennsylvania State University

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 1 / 11


Prerequisite:
MATH 220, MATH 230 or 231, MATH/CMPSC 451 or 455.

Course Brief Description:


This course is a continuation of MATH/CMPSC 451 or 455. We will describe
numerical algorithms for solving nonlinear equations, approximation functions and
data, numerical integration, solving linear systems, Eigenvalue problems, solving
ordinary differential equations and boundary value problems. We will also discuss
the underlying mathematical principles and theories of these numerical methods
and their implementations.

Programming Skills:
Some knowledge of either MATLAB, Octave, Fortran, C, or C++ is strongly
recommended. MATLAB is used for exposition of algorithms during the class.

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 2 / 11


Tentative Outline
Nonlinear Equations



f1 (x1 , x2 , , xn ) = 0
f2 (x1 , x2 , , xn ) = 0

f (x) = 0 = ..


.
fn (x1 , x2 , , xn ) = 0

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 3 / 11


Tentative Outline
Nonlinear Equations



f1 (x1 , x2 , , xn ) = 0
f2 (x1 , x2 , , xn ) = 0

f (x) = 0 = ..


.
fn (x1 , x2 , , xn ) = 0

Newtons Method & Quasi-Newtons Method


Fix Point Iteration & Aitkens Acceleration

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 3 / 11


Tentative Outline
Anpproximation

Approximate data set {(xi , yi )}, i = 0, , n;


Approximate functions f (x).

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 4 / 11


Tentative Outline
Anpproximation

Approximate data set {(xi , yi )}, i = 0, , n;


Approximate functions f (x).

Polynomial Interpolation (Chebyshev Interpolation)


Trigonometric Interpolation (Fast Fourier Transform)
Piecewise Interpolation (Bezier Curves)
Least-squares Method

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 4 / 11


Tentative Outline
Numerical Integration

Z b
f (x)dx =?
a

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 5 / 11


Tentative Outline
Numerical Integration

Z b
f (x)dx =?
a

Interpolatory Quadratures (Gauss-Legendre/Gauss-Legendre-Lobatto)


Adaptive Integration

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 5 / 11


Tentative Outline
Linear Systems



a11 x1 + a12 x2 + + a1n xn = b1
a21 x1 + a22 x2 + + a2n xn = b2

Ax = b ..


.
an1 x1 + an2 x2 + + ann xn = bn

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 6 / 11


Tentative Outline
Linear Systems



a11 x1 + a12 x2 + + a1n xn = b1
a21 x1 + a22 x2 + + a2n xn = b2

Ax = b ..


.
an1 x1 + an2 x2 + + ann xn = bn

Direct Method
Linear Iterative Method (General Stationary Iterative Method)
Krylov Subspace Method (Conjugate Gradient Method and more)
Preconditioning

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 6 / 11


Tentative Outline
Eigenvalue Problems

Given a matrix A, find a scalar and x 6= 0, such that:

Ax = x

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 7 / 11


Tentative Outline
Eigenvalue Problems

Given a matrix A, find a scalar and x 6= 0, such that:

Ax = x

Power Method
Inverse Method
QR Method

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 7 / 11


Tentative Outline
Ordinary Differential Equations

0

y1 = f1 (t, y1 , y2 , , yn )
y 0 = f2 (t, y1 , y2 , , yn )

2
y 0 (t) = f (t, y (t)) = .. .. ..

. . .
0

yn = fn (t, y1 , y2 , , yn )

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 8 / 11


Tentative Outline
Ordinary Differential Equations

0

y1 = f1 (t, y1 , y2 , , yn )
y 0 = f2 (t, y1 , y2 , , yn )

2
y 0 (t) = f (t, y (t)) = .. .. ..

. . .
0

yn = fn (t, y1 , y2 , , yn )

Euler Method & Crank-Nicolson Method


Stability Issue
High Order Method (Runge-Kutta Method and Multistep Method)
Predictor-corrector Method
System of Equations

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 8 / 11


Tentative Outline
Boundary Value Problems

Poisson Equation

u 00 (x) = f (x), x (a, b), and u(a) = ua , u(b) = ub

Convection Diffusion Equation

u 00 (x) + b(x)u 0 (x) = f (x), x (a, b), and u(a) = ua , u(b) = ub

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 9 / 11


Tentative Outline
Boundary Value Problems

Poisson Equation

u 00 (x) = f (x), x (a, b), and u(a) = ua , u(b) = ub

Convection Diffusion Equation

u 00 (x) + b(x)u 0 (x) = f (x), x (a, b), and u(a) = ua , u(b) = ub

Finite Difference Method for 1-D Poisson Problem


Finite Difference Method for 1-D Convection-diffusion Problem
Consistency & Convergence of Finite Difference Method
Finite Difference Method for 2-D Problems

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 9 / 11


Recommended reference:
Numerical Analysis, Timothy Sauer, published by Addison Wesley, ISBN
0-321-26898-9.

Grading Policy:
1 Homework & Computer projects (50 %);
2 Midterm exam (20 %): March 2;
3 Final exam (30 %).
Details about midterm and final exam will be announced later.

Office Hours:
Tuesday & Thursday, 1:00 pm - 2:00 pm, or by appointment.

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 10 / 11


Questions?

X. Hu (Penn State) MATH/CMPSC 456 Jan. 9, 2012 11 / 11

Вам также может понравиться