The standard answer is to replace each x, y and z with the corresponding element of the input vector:
g = @(V) = V(1).^2 + 3.*V(2).^2 + 4.*V(3).^2 - 2.*V(1).*V(2) + 5.*V(1) - 3.*V(2) + 2.*V(3);
A more readable alternative is to use two functions, which allows you to keep the original function unchanged. This could be useful if it is important to keep exactly the same equation as some reference, or to use exactly those variables names, or just to maintain readability:
f = @(x,y,z) = x.^2 + 3.*y.^2 + 4.*z.^2 - 2.*x.*y + 5.*x - 3.*y + 2.*z;
g = @(V) f(V(1),V(2),V(3));
and then call the function g in all of your following work. It has the disadvantage that it would likely be a smidgen slower, so do not use this if your need to perform millions of iterations.
