Let $\Vert x\Vert = \max\Big\{\frac{|x_1|}{3}, \frac{|x_2|}{2}\Big\}$ with $x \in \mathbb{R}^2$. Show that $\Vert\cdot\Vert$ is a norm on $\mathbb{R}^2$.

I do have problems showing that the triangle inequality holds.

So far, I tried the following: Let $x, y \in \mathbb{R}^2$, than

$$ \begin{align*} \Vert x\Vert &= \max\Bigg\{\frac{|x_1|}{3}, \frac{|x_2|}{2}\Bigg\}\\ &= \frac{1}{2}\left(\frac{|x_1|}{3} + \frac{|x_2|}{2} + \Bigg|\frac{|x_1|}{3} - \frac{|x_2|}{2}\Bigg|\right)\\ &=\frac{1}{6}\left( \frac{1}{2}\left(2|x_1| + 3|x_2| + \Big|2|x_1| -3|x_2|\Big|\right)\right)\\ &=\frac{1}{6}\max\Big\{2|x_1|, 3|x_2|\Big\}, \end{align*} $$


$$\small \begin{align*} \Vert x + y \Vert &\leq \Vert x \Vert + \Vert y \Vert\\ \Leftrightarrow \max\Bigg\{\frac{|x_1 + y_1|}{3}, \frac{|x_2 + y_2|}{2}\Bigg\} &\leq \max\Bigg\{\frac{|x_1|}{3}, \frac{|x_2|}{2}\Bigg\} + \max\Bigg\{\frac{|y_1|}{3}, \frac{|y_2|}{2}\Bigg\}\\ \Leftrightarrow \frac{1}{6}\max\Big\{2|x_1 + y_1|, 3|x_2+y_2|\Big\} &\leq \frac{1}{6}\left(\max\Big\{2|x_1|, 3|x_2|\Big\} + \max\Big\{2|y_1|, 3|y_2|\Big\}\right)\\ \Leftrightarrow 2|x_1+y_1| + 3|x_2+y_2| + \Big|2|x_1+y_1| -3|x_2+y_2|\Big| &\leq 2|x_1| + 3|x_2| + \Big|2|x_1| -3|x_2|\Big| + 2|y_1| + 3|y_2| + \Big|2|y_1| -3|y_2|\Big|\\ \Leftrightarrow \underbrace{2|x_1+y_1| + 3|x_2+y_2|}_{\leq\left(2|x_1| +2|x_2|\right) + \left(3|x_1| + 3|x_2|\right)} + \Big|2|x_1+y_1| -3|x_2+y_2|\Big| &\leq \left(2|x_1| +2|x_2|\right) + \left(3|x_1| + 3|x_2|\right) + \Big(\Big|2|x_1| -3|x_2|\Big| + \Big|2|y_1| -3|y_2|\Big|\Big) \end{align*} $$

I have problems showing the last inequality. I do know that $\Big|2|x_1+y_1| -3|x_2+y_2|\Big| \leq \Big|2|x_1| -3|x_2|\Big| + \Big|2|y_1| -3|y_2|\Big|$ is not true in general. So I suppose there is something I do not see?

  • $\begingroup$ An alternative solution is to notice that this is the well-known $\ell^\infty = \|-\|_\infty$ norm on $T(\mathbb{R}^2)$, where $T$ is the diagonal matrix with $t_{11} = 1/3$, $t_{22} = 1/2$. This greatly simplifies the proof: $$\|T(x+y)\|_\infty = \|Tx + Ty\|_\infty \leqslant \|Tx\|_\infty + \|Ty\|_\infty$$ by the fact that we know $\|-\|_\infty$ is a norm. $\endgroup$ – William May 10 at 3:47

To prove that the triangles inequality holds, we need to show that $$\|x+y\|\leq \|x\|+\|x\|$$ $$\Leftrightarrow \max\Bigg\{\frac{|x_1 + y_1|}{3}, \frac{|x_2 + y_2|}{2}\Bigg\} \leq \max\Bigg\{\frac{|x_1|}{3}, \frac{|x_2|}{2}\Bigg\} + \max\Bigg\{\frac{|y_1|}{3}, \frac{|y_2|}{2}\Bigg\} $$ which is equivalent to $$\frac{|x_1 + y_1|}{3} \leq \max\Bigg\{\frac{|x_1|}{3}, \frac{|x_2|}{2}\Bigg\} + \max\Bigg\{\frac{|y_1|}{3}, \frac{|y_2|}{2}\Bigg\} $$ and $$\frac{|x_2 + y_2|}{2} \leq \max\Bigg\{\frac{|x_1|}{3}, \frac{|x_2|}{2}\Bigg\} + \max\Bigg\{\frac{|y_1|}{3}, \frac{|y_2|}{2}\Bigg\} $$ Note that $$\frac{|x_1 + y_1|}{3} \leq \frac{|x_1|}{3}+\frac{|y_1|}{3}\leq\max\Bigg\{\frac{|x_1|}{3}, \frac{|x_2|}{2}\Bigg\} + \max\Bigg\{\frac{|y_1|}{3}, \frac{|y_2|}{2}\Bigg\}$$ The second inequality can be obtained similarly.

  • 1
    $\begingroup$ +1 The OP should look t this, see how much clearer it is than his attempt, and learn from it. $\endgroup$ – GEdgar Apr 4 at 11:01
  • $\begingroup$ @GEdgar thank you! $\endgroup$ – JwJJJJ Apr 5 at 1:20

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.