1
$\begingroup$

Let $A$ be a symmetric matrix and $X$ a symmetric positive definite matrix, then the following standard semidefinite optimization problem is convex:

min $tr (AX)$ subject to $X>0$

Now I wonder if $tr ((-A)X)$ is convex or concave?

The reason is the following. I want to solve: max $tr (AX)$ subject to $X>0$

Can that be written as: min $tr ((-A)X)$ subject to $X>0$ ? Is that convex and hence solvable?

$\endgroup$
  • 1
    $\begingroup$ Any linear function of $X$ is convex. $\endgroup$ – littleO Oct 11 '13 at 11:02
1
$\begingroup$

It is still convex. Note that for semi-definite optimization, the requirement is that the objective function and constraints (except the semi-definite constraint) are linear in terms of the entries of $\mathbf{X}$.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.