Considering the following optimization program:

$$
   maximize \ \ \ \log \left( \|x\|_\infty \right)
$$

$$
subject \ to \ \ \ \  Ax\leq b
$$

can we rewrite this program as a convex equivalence?