4

$V\subset \mathbb{R}^n $ be open, symmetric and contains 0 . Define $||\cdot||:\mathbb{R}^n \rightarrow \mathbb{R}$ as:

\begin{equation} ||x||=\inf\{ \lambda > 0: \lambda^{-1}x\in V \} \end{equation} for all x $\in \mathbb{R}^n$.

Prove this is a norm $\iff$ $V$ is convex and bounded with respect to the euclidean metric.

Approach:

  1. $||x||\ge 0$ (Done)
  2. $||\alpha x||= |\alpha|||x||$ (not done)
  3. triangle in equality (not done)
mmcrjx
  • 664

1 Answers1

1

I guess it depends what you mean by symmetric? Inversion symmetric (i.e. contains $x$ iff it contains $-x$)? This is not a norm depending on what you mean. Consider the open set $(-10,10)\times (-1,1)$. This is inversion symmetric, contains the origin, etc. Let $z_1=(1/\sqrt{2},1/\sqrt{2})$, and $z_2=(1/\sqrt{2},-1/\sqrt{2})$. We have:

\begin{align} \|z_1\| &= \|z_2\| = 2\\ \|z_1+z_2\| &= 10/\sqrt{2} \end{align}

Property 2 appears to hold though: That is, this functional is homogeneous.