26 lines
610 B
Matlab
26 lines
610 B
Matlab
|
function [r, J] = jacobian(fun, x, h)
|
|||
|
|
|
|||
|
|
% Returns the Jacobian matrix associated to a function.
|
|||
|
|
%
|
|||
|
|
% INPUTS:
|
|||
|
|
% - fun [handle] Function from Rᵖ to Rⁿ.
|
|||
|
|
% - x [double] 1×p vector, point where the Jacobian has to be evaluated.
|
|||
|
|
% - h [double] scalar, perturbation size.
|
|||
|
|
%
|
|||
|
|
% OUTPUTS:
|
|||
|
|
% - r [double] n×1 vector, value of the function at point x.
|
|||
|
|
% - J [double] n×p matrix of derivatives.
|
|||
|
|
|
|||
|
|
r = fun(x);
|
|||
|
|
|
|||
|
|
if nargout>1
|
|||
|
|
z = x;
|
|||
|
|
n = length(r);
|
|||
|
|
p = length(x);
|
|||
|
|
J = zeros(n, p);
|
|||
|
|
for i=1:p
|
|||
|
|
x = z;
|
|||
|
|
x(i) = x(i)+h;
|
|||
|
|
J(:,i) = (r-fun(x))/h;
|
|||
|
|
end
|
|||
|
|
end
|