4
$\begingroup$

I've been trying to prove that the following is a norm, but wasn't successful. I also cannot find a counterexample. So help is greatly appreciated. Let $x \in \mathbb{R}^N, \ w_i \in \mathbb{R}_+,\ i=1,\ldots,N$. $$\lVert x \rVert_w := \max \lvert w_i x_i\rvert$$

Basically, this is the maximum norm with positive weights assigned to each dimension.

It must be shown that: $$\max \lvert w_i (x_i+y_i) \rvert \leq \max \lvert w_j x_j \rvert + \max \lvert w_k y_k \rvert$$

$\endgroup$
  • $\begingroup$ If you assume that $\mathbb{R}_+=[0,+\infty)$ then $\Vert x\Vert_w$ is not a norm, just a seminorm. $\endgroup$ – Norbert Aug 13 '12 at 13:16
  • $\begingroup$ @Norbert: (quote) with positive weights assigned to each dimension. $\endgroup$ – Did Aug 13 '12 at 13:25
4
$\begingroup$

Since $|w_ix_i|\leqslant \lVert x \rVert_w$ and $|w_iy_i|\leqslant \lVert y \rVert_w$ for every $i$, $$ |w_i(x_i+y_i)|\leqslant|w_ix_i|+|w_iy_i|\leqslant\lVert x\rVert_w+ \lVert y \rVert_w, $$ for every $i$. This proves that $\max\limits_i|w_i(x_i+y_i)|\leqslant\lVert x\rVert_w+ \lVert y \rVert_w$, QED.

$\endgroup$

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.