Mansoureh
Oct20-11, 02:15 AM
Hi,
I applied Gauss-lobatto method to solve elliptic integral in my problem. one of my parameters(distance) is infinity. in my problem the answer of elliptic integral at large distance should be zero but I can't get this result numerically. when distance goes to infinity, my program stopped working.
I am looking for the resean and I guess this might be because the Gauss-lobatto method is not accurate for my problem.
I used following code:
N = max(3*round((N−1)/3),3) + 1; % Adjust N to the closest valid choice
h = (b − a)/(N−1);
d = (3/sqrt(5) − 1)*h/2;
x = (a:h:b).'; x(2:3:N−2) = x(2:3:N−2) − d; x(3:3:N−1) = x(3:3:N−1) + d;
w = ones(1,N); w(4:3:N−3) = 2; w([2:3:N−2,3:3:N−1]) = 5; w = w*h/4;
I = w * f(x); % Approximately evaluate the integral
please let me know if you have any idea.
Thanks,
I applied Gauss-lobatto method to solve elliptic integral in my problem. one of my parameters(distance) is infinity. in my problem the answer of elliptic integral at large distance should be zero but I can't get this result numerically. when distance goes to infinity, my program stopped working.
I am looking for the resean and I guess this might be because the Gauss-lobatto method is not accurate for my problem.
I used following code:
N = max(3*round((N−1)/3),3) + 1; % Adjust N to the closest valid choice
h = (b − a)/(N−1);
d = (3/sqrt(5) − 1)*h/2;
x = (a:h:b).'; x(2:3:N−2) = x(2:3:N−2) − d; x(3:3:N−1) = x(3:3:N−1) + d;
w = ones(1,N); w(4:3:N−3) = 2; w([2:3:N−2,3:3:N−1]) = 5; w = w*h/4;
I = w * f(x); % Approximately evaluate the integral
please let me know if you have any idea.
Thanks,