I have the following example in my lectures notes:
In $(\mathbb{R}^n,\varepsilon_n)$ , where $\varepsilon_n$ is the usual euclidean topology or the one induced from it. Let $ B=B(O,1)$ be the open ball with center the origin $O$ and radious $1$ and $Q=\{x \in \mathbb{R}^n| | x_i| <1, i=1,...,n \}$ the n-dimensional cube of center $O$ and side $ 2 $ (or in other words the open ball of center $O$ and radious $r$ with rispect to the metric $d_{L^{\infty}}$) We want to prove $ B$ is homeomorphic to $Q$

Let $x=(x_1,...,x_n) \in \mathbb{R}$, we have that $\|x\|_{L^{\infty}}=d_{L^{\infty}}(x,O)=max\{|x_i|| i=1,...,n\}$. Let $k(x)=\frac{\|x\|_{L^{\infty}}}{\|x\|}$, and let's consider the mapping $f:(\mathbb{R}^n,\varepsilon_n) \rightarrow (\mathbb{R}^n,\varepsilon_n)$ defined by: $$f(x)=\begin{cases} k(x)x & \text{if } x \neq 0, \\ 0 &\text{if } x=0\end{cases}$$

Now, $f$ is a homeomorphism from $(\mathbb{R}^n,\varepsilon_n)$ to itself such that $f(Q)=B$, that is $f$ is open

In fact $ 1/\sqrt(n) \leq k(x) \leq 1$ and $k(\alpha x)=k(x)$ for each $x \neq O$ and $ \alpha \in \mathbb{R}^n\setminus\{0\}$, then $ \lim_{x \to O}f(x)=O$ and $f$ is continuous over $\mathbb{R}^n$

Furthermore the mapping $$g(x)=\begin{cases} x\frac{1}{k(x)} & \text{if } x \neq 0, \\ 0 &\text{if } x=0\end{cases}$$

is defined and continous over $\mathbb{R}^n$ and $g\circ f=f \circ g = I_{\mathbb{R}^n}$, then $g=f^{-1}$ and f is a homeomorphism. Besides $\|f(x)\|=\|x\|_{L^{\infty}}$ and $\|g(x)\|_{L^{\infty}}=\|x\|$, so $f(Q)=B$

I need some help filling in the details . To prove f is an homeomorphism I have to prove it is continuous, bijective and open.

(1) continuity.

The relation $ 1/\sqrt(n) \leq k(x) \leq 1$ comes from the equivalence of the $L^p$ metrics:

I know $\|x\|_{L^{\infty}} \leq \|x\|_{L^p} \leq n^{1/p}\|x\|_{L^{\infty}}$ for $p=2$ and $\|x\|_{L^2}=\|x\|$ I get $ 1/\sqrt(n) \leq k(x) \leq 1$. Also, the expresion $k(\alpha x)=k(x)$ is straightforward, but then I can't figure out how to show $ \lim_{x \to O}f(x)=O$

(2)injectivity For the injectivity I tried the ususal approach: $f(x_1)=f(x_2)$, so $\frac{\|x_1\|_{L^{\infty}}x_1}{\|x_1\|} = \frac{\|x_2\|_{L^{\infty}}x_2}{\|x_2\|}$ but then I don't know how to get to $x_1=x_2$

(3) surjectivity For the surjectivity I need to show the equation $\frac{\|x\|_{L^{\infty}}x}{\|x\|}=y$ has a solution for any $y \in \mathbb{R}^n$, but I can't figure out how to solve for $x$


I don't get how $\|f(x)\|=\|x\|_{L^{\infty}}$ and $\|g(x)\|_{L^{\infty}}=\|x\|$ manage to conclude $f(Q)=B$

I was trying to do like this: By definition of image $f(Q)=\{y|f(x)=y, x \in Q\}$ I take $x \in Q \iff\|x\|_{L^{\infty}}<1 \iff$ $\|x\|_{L^{\infty}}= \|\frac{1}{k(x)}y\|_{L^{\infty}}=\frac{1}{k(x)}\|y\|_{L^{\infty}}<1 $ $\implies \|y\|_{L^{\infty}}<k(x)\leq 1$ and I am getting nowhere, I should be getting to something like $\|y\|<1$ but instead I have the infinity norm there

Any help would be much appreciated

  • $\begingroup$ What is $\varepsilon_n$ in $(\mathbb R^n,\varepsilon_n)$ in your problem statement? $\endgroup$ – kimchi lover Jul 7 '20 at 15:18
  • $\begingroup$ @kimchi lover Its the usual euclidean topology $\endgroup$ – J.C.VegaO Jul 7 '20 at 15:20
  • $\begingroup$ Thanks. Edit your post to say so. You have $\varepsilon_n, \|\cdot\|$ and $\|\cdot\|_{L^2}$, all meaning the same thing. May I suggest you replace them all with $\|\cdot\|_2$, and so on? $\endgroup$ – kimchi lover Jul 7 '20 at 15:21
  • $\begingroup$ @kimchi lover all right, $\varepsilon_n$ is the euclidean topology. I think others denote like $\tau_e$, for other two you are right $\endgroup$ – J.C.VegaO Jul 7 '20 at 15:27
  • $\begingroup$ @kimchi lover Actually , I did it on purpose to see the conection of $p$ with the inequality in which it appears as an exponent: $\|x\|_{L^{\infty}} \leq \|x\|_{L^p} \leq n^{1/p}\|x\|_{L^{\infty}}$, then I did put $\|x\|_{L^2}=\|x\|$, $\endgroup$ – J.C.VegaO Jul 7 '20 at 15:29


$\lim_{x\rightarrow 0}\|k(x)x\|=\lim_{x\rightarrow 0} \frac{\|x\| \ \|x\|_{L^\infty}}{\|x\|}=0$

Injectivity and surjectivity

I would suggest the following aproach:

Prove that: $x\in Q$ implies that $f(x)\in B$ $x\in B$ implies that $g(x)\in Q$

Then prove that $g$ is the inverse of $f$, i.e. that $f(g(x))=x$ and $g(f(x))=x$.

This is enough to show that $f$ is a bijection from $Q$ to $B$.

  • $\begingroup$ Not sure how $x\in Q$ implies that $f(x)\in B$ , $x\in B$ implies that $g(x)\in Q$ together wih having the inverse implies f is bijective $\endgroup$ – J.C.VegaO Jul 7 '20 at 17:51
  • $\begingroup$ Functions with an inverse are bijective. math.stackexchange.com/q/1215365/15624 $\endgroup$ – Angela Pretorius Jul 8 '20 at 4:04
  • $\begingroup$ ok, and what about these: " $x\in Q$ implies that $f(x)\in B$ , $x\in B$ implies that $g(x)\in Q$" what are they for? $\endgroup$ – J.C.VegaO Jul 8 '20 at 8:50

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.