0
$\begingroup$

I have a Markowitz problem :

Min $x^T*C*x$

$x : {x_1 , x_2 ... x_n}$ is a vector of size $N$

$C$ is a known matrix $[N \times N]$

1) $∑ x_i$ = 1

2) $x_1 < 0 $

I can minimize the function with the first constraint with Excel Solver. I find the optimal $x$ vector.

It doesn't work with the second one (because of the strict inequality) .

I don't know what kind of problem is this, so if someone have a path of research ..

Thanks

$\endgroup$
  • $\begingroup$ Try for "Quadratic programming". $\endgroup$ – Khue Jun 24 '14 at 9:50
  • $\begingroup$ With the 2nd constraint it seems to be no longer a quadratic problem :s $\endgroup$ – user159854 Jun 24 '14 at 10:05
  • $\begingroup$ The second constraint can be seen as $Ax < 0$ where $A=[1 \ 0 \ \cdots \ 0]$. You can relax it to $Ax \le 0$ and use any quadratic solver. $\endgroup$ – Khue Jun 24 '14 at 10:12
  • $\begingroup$ Ok but the solver will give me x = 0 as solution. I want to keep a strict inequality . $\endgroup$ – user159854 Jun 24 '14 at 11:30
  • $\begingroup$ I guess you forgot to include the first constraint, i.e. $\mathbf{1}^Tx =1$, in the problem to be solved. $\endgroup$ – Khue Jun 24 '14 at 12:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.