1
$\begingroup$

Prove: A first countable hemicompact space is locally compact.

A topological space $(X,\tau)$ is said to be hemicompact if it has a sequence of compact subsets $K_n$, $n \in \mathbb{N}$, such that every compact subset $C$ of $(X,\tau)$ satisfies $C \subseteq K_n$, for some $n \in \mathbb{N}$.

In locally compact space each point $x$ has a compact nbhd $C$, s.t. $x \in U \subseteq C$, where $U$ is open.

By first countability, there exists a monotonic decreasing sequence $U_n$ of open nbhds of point $x$. By hemicompactness, there exists an increasing monotonic sequence $K_m$ of compact sets containing $x$. My intuition tells me that $U_n \subseteq K_m$, for some $n,m \in \mathbb{N}$, and $K_m$ is a compact nbhd of $x$, but I can't formalize this. Am I on the right path?

$\endgroup$
3
  • $\begingroup$ Please could you state the definitions you're using for hemicompact and locally compact? $\endgroup$ – Henno Brandsma Sep 15 '18 at 3:10
  • $\begingroup$ @henno-brandsma I've added the definitions of hemicompact and locally compact $\endgroup$ – Andreo Sep 15 '18 at 17:57
  • $\begingroup$ You're idea is OK, you need a diagonalisation argument to formalise it. See my answer. $\endgroup$ – Henno Brandsma Sep 16 '18 at 15:34
2
$\begingroup$

Let $K_n$ be the cofinal sequence of compact subsets and let $p \in X$ and let $U_n$ be a decreasing (WLOG) local base at $p$.

Suppose that for all $n$: $U_n$ is not a subset of $K_n$. Let $x_n \in U_n \setminus K_n$ be a witnessing point of this non-inclusion.

Then $S = \{x_n: n \in \mathbb{N}\} \cup \{p\}$ is a compact subset of $X$ (a sequence together with its limit), and so by assumption on the cofinal sequence, for some $m$, we must have $S \subseteq K_m$. But this cannot be, as $x_m \in S$ and $x_m \notin K_m$. This contradiction shows that the original assumption is false, so that for some $n$, $U_n \subseteq K_n$, which shows that $K_n$ is a compact neighbourhood of $p$.

As this works for all $p$, $X$ is locally compact, as claimed.

$\endgroup$

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.