两点边界值问题(英文版).pdf

Lecture23 April 22, 2008 Lecture 23: Starting to put it all together #2... More 2-Point Boundary value problems Outline 1) Our basic example again: -u + u = f(x); u(0)=α, u(L)=β 2) Solution of 2-point Boundary value problems A) Shooting Methods B) Finite-Difference methods C) Higher-order: Spectral - Collocation methods D) Variation #2: Non-linear problems E) Galerkin Finite Elements An Example problem: Simplest Steady-state 2-point boundary value problem x=0 x=L u(0,t)=α u(L,t)=βf(x) Lecture23 April 22, 2008 Solving 2-pt ODE Boundary Value Problems (BVPs) Method #2: Finite Difference Schemes Finite Differences and Linear algebra: Boundary conditions Solving 2-pt ODE Boundary Value Problems (BVPs) Method #2: Finite Difference Schemes Matlab implementation: function [ D1, D2, x ] = centerDifference(N,xBounds) % CenterDifference - generates 2nd order sparse Difference matrices and coordinates for % N evenly spaced points in the domain xBounds = [xMin xMax] % % [ D1, D2, x ] = centerDifference(N,xBounds) % % D1: first derivative Matrix 1/2h [ -1 0 1]; % D2: second derivative matrix 1/h^2 [ 1 -2 1 ] % x: uniform grid % % N: number of grid points % xBounds: array holding coordinates of domain [ xMin xMax] % % generate uniform grid x = linspace(xBounds(1),xBounds(2),N); x=x(:); % make colum vector h = x(2)-x(1); % grid spacing % generate sparse Difference matrices e=ones(N,1); D1=spdiags([ -e e], [-1 1],N,N)/2/h; % First derivatve matrix D2=spdiags([ e -2*e e], -1:1,N,N)/h^2; % second derivative matrix % fix boundary points for consistent 1-sided derivatives D1(1,1:3) = [ -3 4 -1 ]/2/h; D1(end,end-2:end) = [ 1 -4 3]/2/h; D2(1,1:3) = [ 1 -2 1]/h/h; D2(end,end-2:end) = [ 1 -2 1]/h/h; Lecture23 April 22, 2008 Solving 2-pt ODE Boundary Value Problems (BVPs) Method #2: Finite Difference Schemes Matlab implementation: function [A,x] = finiteDifferenceOperator(N,xBounds