The gradient is simply the derivative, since yours is a univariate function. I would create a vectorized anonymous function from it, and use it as I would any other function:
dd_dx = @(x) (1303*x)/2500 - (4999*x.^2)/100000 + 25/2;
then call it as:
x = linspace(0, 10);
figure(1)
plot(x, dd_dx(x))
grid
