0
$\begingroup$

I am reading about how a maximization problem can be converted into a minimization problem: what I have understood is the following:

To change a max problem to a min problem, just multiply the objective function by −1.

1) But I am now confused. I noticed some people have considered that $max_{x}{f(x)}$ is equivalent to $-min_x (-f(x))$. So I am wondering why it is not just $min_x (-f(x))$ ? When we should put a negative sign before the $min$? and when we should not?

2) According to 1), consider now the problem $max_Y ||X - Y||_F^2$ (where $X$, $Y$ are two matrices). So what is its equivalent minimization problem? Is it $min_Y -||X-Y||_F^2$ or $-min_Y -||X-Y||_F^2$ ?

Your help is very appreciated.

$\endgroup$
1
$\begingroup$

It should always have the negative sign in front of it because otherwise it's a minimum, not a maximum. Consider the problem of maximizing $f(x)=-x^2 + 1$ for example which has maximum $1$ but $\min_x(-f(x))=-1$ which is not $1$.

$\endgroup$
1
  • $\begingroup$ Thank you very much for the answer $\endgroup$ – Christina Sep 24 '18 at 5:03

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.