Minimizing l2 Norm with gradient descent with multiple variables

Hello, im kind of new to matlab. I want to minimize my l2 error with gradient descent. The l2 error depends on two variables g,h which i want to initialize at 0.5 , 0.5. Can someone help me? How do i put my l2-error and steepest gradient descent together?

[Torsten]

Differentiate the l2-error with respect to g and h.

Then -gradient(l2_error) will be the descent direction for the method of steepest descent.

As soon as the distance between two subsequent values X_i = (g_i,h_i) and X_(i+1) = (g_(i+1),h_(i+1)) coming from the formula

X_(i+1) = X_(i) - r*gradient(l2_error)

for a constant r>0 is less than a small number eps chosen by you, stop the algorithm.