4
$\begingroup$

Recently I came to know that for $n\geq 2$, the open ball $B(0,1) = \{z\in \mathbb{C}^n: |z|<1\}$ and the polydisc $\Delta(0;1) = \{z\in \mathbb{C}^n: |z_j|<1, j=1,\ldots, n\}$ are not biholomorphic.

This made me wonder whether $B(0,1) = \{x\in \mathbb{R}^n: ||x||<1\}$ and the polydisc $\Delta(0;1) = \{x\in \mathbb{R}^n: |x_j|<1, j=1,\ldots, n\}$ are diffeomorphic?

I know that the product topology on $\mathbb{R}^n$ is equivalent to the metric topology, i.e. topology generated by open polydiscs is same as the one generated by open balls. Does this imply that $B(0,1)$ is homeomorphic to $\Delta(0;1)$?

$\endgroup$
1
  • $\begingroup$ @Dap Wouldn't it be worth to write an official answer even it is just a link? $\endgroup$
    – Paul Frost
    Feb 26, 2019 at 12:41

3 Answers 3

4
+50
$\begingroup$

I know that the product topology on $\mathbb{R}^n$ is equivalent to the metric topology, i.e. topology generated by open polydiscs is same as the one generated by open balls. Does this imply that $B(0,1)$ is homeomorphic to $\Delta(0;1)$?

Not if it's for a topological reason. For example, the topology on $\mathbb R^n$ is also generated by the sets which are unions of two open balls of radius $r$ and whose centers are at distance $5r$. But a disjoint union $B(0, 1) \sqcup B(0, 1)$ is definitely not homeomorphic to $B(0, 1)$.

  • We have a smooth map $$\begin{align*} B(0, 1) & \to \mathbb R^n \\ x & \mapsto \frac{x}{\sqrt{1- \|x\|^2}} \end{align*}$$ with smooth inverse $$\begin{align*} \mathbb R^n & \to B(0, 1) \\ y & \mapsto \frac{y}{\sqrt{1 + \|y\|^2}} \end{align*}$$

  • If $g : (-1, 1) \to \mathbb R$ is any diffeomorphism, then $$\begin{align*} \Delta(0, 1) & \to \mathbb R^n \\ (x_1, \ldots, x_n) & \mapsto (g(x_1), \ldots, g(x_n)) \end{align*}$$ is a smooth map with smooth inverse $$\begin{align*} \mathbb R^n & \to \Delta(0, 1) \\ (y_1, \ldots, y_n) & \mapsto (g^{-1}(y_1), \ldots, g^{-1}(y_n)) \end{align*}$$

$\endgroup$
3
$\begingroup$

They are diffeomorphic since they are both open star shaped sets, see A star shaped open set in $\mathbb{R}^n$ is diffeomorphic to $\mathbb{R}^n$

$\endgroup$
3
  • $\begingroup$ How to prove that a polydisc is star-shaped? I expected that there will be a direct proof of diffeomorphism; similar to the proof of open ball being diffeomorphic to Euclidean space. Also, with my present knowledge level, I couldn't understand the proof given in the link. I will be grateful if you could also give a brief outline of the proof. $\endgroup$ Feb 26, 2019 at 16:49
  • $\begingroup$ @rationalbeing: it's star shaped at zero because $z\in \Delta(0;1)$ implies $\lambda z\in\Delta(0,1)$ for $\lambda\in[0,1].$ The answer by "40 votes" is quite conceptual and just relies on an intuitive fact about ODEs. $\endgroup$
    – Dap
    Feb 26, 2019 at 18:57
  • $\begingroup$ And yes there is probably a much simpler proof for this case $\endgroup$
    – Dap
    Feb 26, 2019 at 19:02
3
$\begingroup$

You can indeed show that two balls for two (equivalent) norms are homeomorphic :

Homeomorphism between open unit balls from equivalent distances

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .