$G$ is a locally compact (may not necessarily Hausdorff) group, $H$ is a subgroup in $G$, $G/H$ is compact as a quotient space , then there exist a compact subset $K$ such that $G=KH$(or $G=HK$).

  • $\begingroup$ is G/H assumed to be hausdorff? $\endgroup$ – oxeimon Jun 10 '15 at 18:25
  • $\begingroup$ Use that $G/H$ is compact. $\endgroup$ – Dietrich Burde Jun 10 '15 at 18:34
  • $\begingroup$ @oxeimon: $G/H$ may not necessarily be Hausdorff. $\endgroup$ – David Chan Jun 10 '15 at 18:42
  • $\begingroup$ Write $G/H=K$. Then G=KH$. $\endgroup$ – Dietrich Burde Jun 10 '15 at 18:44
  • $\begingroup$ Use this to write it down. $\endgroup$ – Dietrich Burde Jun 10 '15 at 19:09

It seems to me that what you want follows from the result of the following paper:


The above paper show the existence of an interesting cross section $q$ of the fibration $\pi: G \to G/H$. Namely, a map $q: G/H \to G$ such that $\pi \circ q = id_{G/H}$, with the additional property that $q(C)$ is a relatively compact subset of $G$ for any compact subset $C \subset G/H$. So, if you assume $G/H$ to be compact then by setting $$ K := \overline{q(G/H)}$$ you get a compact subset $K \subset G$ such that $G = K H$ as the OP asked.


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.