Defining constraints in the Quadratic Programming

Hi,

I have a question to the quadprog command of matlab. Is it possible to declare single constraints?? For example I have x1,x2,x3 and I know the value of the x1 and x3. Now, I need the optimal value for the x2.

Thanks ahead,

Philipp

Answer 1

[Alan Weiss]

You can try setting equal upper and lower bounds for those variables. For example,

lb = [3, -inf,5];
ub = [3,inf,5];

sets x(1) = 3 and x(3) = 5, while leaving x(2) unconstrained.

Of course, it might be even better to just reformulate your problem to have fewer variables, but I know that is sometimes difficult.

Alan Weiss

MATLAB mathematical toolbox documentation

Answer 2

[Matt J]

If you have a quadratic function of only 1 unknown variable, it should be possible to find the minimum analytically - no iterative optimization required.