Gradient Descent

This function finds the local minima of a function.

Finds local minima according to http://en.wikipedia.org/wiki/Gradient_descent

Run test_grad_descent for example.

>>The first output of the function should be the function value whereas the
>>second should be the gradient

% Inputs:
% func : the function in which to be optimised over, must
% return value as first argument and gradient as second
% x : starting point to begin optimisation
% gamma (optional) : minimising parameter, small values will get to the
% local maxima, slowly whereas bigger parameters will
% be faster but might overshoot (default 0.2)
% iter (optional) : Number of iterations over which it is optimised
% (default 100)
% thresh (optional): stopping criteria for which the function is not
% minimised any more. (default 1e-6)
% display (optional): Whether or not to display ouput of iteration,
% function value, and change (default 1)
%
% Outputs:
% x: locally optimised location
% val: f(x) i.e. the function value

Thanks,
Sachin