MathOverflow will be down for maintenance for approximately 3 hours, starting Monday evening (06/24/2013) at approximately 9:00 PM Eastern time (UTC-4).

## Picking a representative in a continuous way

I'm hoping for some ideas/pointers here. I'm experimenting with a Livschitz theorem for functions on a locally compact Abelian group, where the periodic orbit sums take values in a closed subgroup.

If $G$ is a locally compact Abelian group (second countable) and $H$ is a closed subgroup then $G$ and $H$ inherit a natural translation-invariant metric. You can use this to induce a translation-invariant metric on $G/H$.

Does there exist a neighbourhood $N$ of the identity in $G/H$ and a continuous map $f\colon G/H\to G$ such that $f(x)\in x$ for all $x\in N$?

EDITED to incorporate simplifying suggestion of Emil Jerabek and correction of Mark Schwarzmann.

-
 WLOG $X=G/H$, $f=\operatorname{id}$, and $x=0$. – Emil Jeřábek Apr 12 2011 at 16:06 I agree that $X=G/H$ and $x=0$ wlog. Not sure what you mean by $f=id$ though. $f$ is supposed to map $G/H$ into $G$. – Anthony Quas Apr 12 2011 at 16:42 It does now, but in the original version of the question which I was commenting on, $f$ was a function from $X$ to $G/H$, and what you now call $f$ was called $F$. – Emil Jeřábek Apr 12 2011 at 17:14 What do you mean by $x \in f(x) + H$ for all $x \in N$? $x$ is a coset of $H$, what does it mean for it to belong to another coset of $H$? – Mark Schwarzmann Apr 12 2011 at 18:49

You are essentially asking for a local cross section for the fiber bundle $(G,\pi: G \to G/H)$. According to this article, it is known that such a section exists if the group $G$ is Lie, or more generally, LCSC and with finite covering dimension. I'm not an expert on the subject, but it seems to me that there exist metrizable LCSC abelian groups with infinite covering dimension, so this doesn't seem to give a complete answer to your question. Perhaps someone who knows more on the subject can illuminate this point, or provide a more recent reference.