Matlab Gauss quadrature

% matlab script to demonstrate use of Gauss quadrature
 
clear all
close all
 
% first derive the 2-point Gauss quadrature rule
 
eq1 = 'w1*1 + w2*1 = 2';
 
eq2 = 'x1*w1 + x2*w2 = 0';
 
eq3 = 'x1^2*w1 + x2^2*w2 = 2/3';
 
eq4 = 'x1^3*w1 + x2^3*w2 = 0';
 
[w1,w2,x1,x2] = solve(eq1,eq2,eq3,eq4)
pause
 
% note: there are two solutions
% we pick the one where x1 < x2
 
[x1,index] = min(double(x1))
w1 = double(w1(index))
 
x2 = double(x2(index))
w2 = double(w2(index))
 
% define the integrand
 
f = @(t) exp(-((t+1)./2).^2)
 
% use Gauss-Legendre quadrature to approximate it
 
gq = (w1*feval(f,x1) + w2*feval(f,x2));
% don't forget the scaling!
gq = gq/2
 
% now let's find the exact answer and see what the error is ...
 
syms x
I = double(int(exp(-x^2),0,1))
 
errorGQ = I - gq
 
% compare to Simpson's rule
 
a = -1; b = 1; c = (a+b)/2;
s = (b-a)/6*(feval(f,a)+4*feval(f,c)+feval(f,b));
% don't forget the scaling!
s = s/2
 
errorS = I - s