Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

$L\subset \mathbb{R}^{n}$ lattice of max. rank

$\Leftrightarrow \mathbb{R}^{n}/L$ compact in quotient topology

$\Leftrightarrow \exists$ bounded subset $B\subset\mathbb{R}^{n}$ s.t. $L+B = \mathbb{R}^{n}$

I just managed to do the first implication:

$L = \mathbb{Z} \omega_{1} + ... + \mathbb{Z} \omega_{n}$.
$P:=\{\lambda_{1}\omega_{1} + ...+ \lambda_{n}\omega_{n} | 0 \leq \lambda_{i} \leq 1\}.$
Then there exists an equivalence relation on $\mathbb{R}^{n}$, s.t. two numbers $x,x'$ are equivalent modulo $L$ $\Leftrightarrow$ $x-x'\in L$. $\pi:\mathbb{R}^{n}\rightarrow\mathbb{R}^n/L$ with $x\mapsto x+L$. We can find for all $x\in\mathbb{R}^{n}$ a $x'\in P$ s.t. $x-x'\in L$.
$\Rightarrow \pi(x)=\pi(x')$ $\Rightarrow \pi\big|_{P} \rightarrow \mathbb{R}^n$/L bijective.
Hence $\pi(P)=\mathbb{R}^n/L$ and since $P$ is compact, so is $\mathbb{R}^n/L$

For the next implication I have no idea yet.
I guess I have to use that $\mathbb{R}^n = \bigcup\limits_{i=1}^n U_{i}$, but I don't know how to get to $L+B = \mathbb{R}^{n}$. Maybe I must also use that $L$ is discrete? But do I only know that from the lattice, or does my proof for the first implication show that as well?

Hope someone feels like helping :)

best, Kathrin

share|cite|improve this question
up vote 0 down vote accepted

(ii) $\to$ (iii): Let $B_\epsilon(x)$ denote the open $\epsilon$-ball around $x$. Denote by $\pi \colon \mathbb R^n\to \mathbb R^n/L$ the projection. As $\mathbb R^n/L$ is compact, and $\pi$ is open, there is a finite set $I \subseteq\mathbb R^n$ such that $\mathbb R^n \subseteq \bigcup_{i\in I} B_\epsilon(x_i)$, taking preimages under $\pi$, we get $$ \mathbb R^n \subseteq \bigcup_{i\in I} B_\epsilon(x_i) + L $$ and $B := \bigcup_{i\in I} B_\epsilon(x_i)$ is bounded.

(iii) $\to$ (i): If $L$ had not full rank, there were an $x \in \mathbb R^n\setminus \{0\}$ such that $x \perp L$ (we may assume that $0 \in L$). So foreach $\lambda \in \mathbb R$ and each $l \in L$: $$ \|\lambda x - l\| = \bigl(\lambda^2 \|x\|^2 + \|l\|^2\bigr)^{1/2} \ge |\lambda| \|x\| $$ If $B$ is bounded by $M$ say, choose $\lambda$ so large that $\lambda\|x\| > M$, then $$ \|\lambda x - (l+b)\| \ge \|\lambda|\|x\| - M > 0 $$ for each $l \in L$, $b \in B$. So $\lambda x \not\in L+B$.

share|cite|improve this answer
Thanks a lot!! :) – Kathrin Nov 22 '12 at 1:21

HINT: Although it’s efficient to prove a ring of implications, in this case that may not be the approach that offers the most insight into what’s really going on.

After proving $(i)\to(ii)$, you might consider proving $(ii)\to(i)$. If $L$ is generated by $\omega_1,\dots,\omega_m$ with $m<n$, choose $\omega\in\Bbb R^n\setminus\operatorname{span}\{\omega_1,\dots,\omega_m\}$, and show that $\pi\upharpoonright \Bbb R\omega$ is injective and hence that $R/L$ is not compact. For $(i)\to(iii)$ just observe that $\Bbb R^n=L+P$. Finally, the same idea that I suggested for $(ii)\to(i)$ can be used to show $(iii)\to(i)$ (or rather, to be precise, the contrapositive $\lnot(i)\to\lnot(iii)$).

share|cite|improve this answer
ah, nice. Thank you as well!! I'll have a look at this! – Kathrin Nov 22 '12 at 1:23
@Kathrin: You’re welcome. – Brian M. Scott Nov 22 '12 at 1:26

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.