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

Proposition Let $(M,d)$ be a metric space. If $K\subset M$ is compact, then $K$ is complete.

Proof Let $\{x_n\}_{n=1}^\infty \subset K$ be a Cauchy sequence, then $$ \forall \varepsilon > 0 \quad \exists n_0 \in \mathbb N \quad \forall m,n\geq n_0 \quad d(x_n,x_m) < \varepsilon. $$ $(\text{Step } 1)$ Since $K$ is compact, $\{x_n\}_{n=1}^\infty$ has an accumulation point $x^*\in K$ (already proved for every metric space). Let's see that $\lim_{n\to\infty} x_n = x^*$.

$(\text{Step } 2)$ Since $x^*$ is an accumulation point $$\forall k\in \mathbb N \quad \exists x_{n_k} \quad\text{s.t.}\quad d(x_{n_k}, x^*) <\frac{1}{k} $$

$(\text{Step } 3)$ Let $\varepsilon > 0$, then $\exists k_0 \quad n_{k_0} \geq n_0 \quad d(x_{n_{k_0}} x^*) < \varepsilon $. Now if $n \geq n_0$, then $$ d(x_n,x*) \leq d(x_n, x_{n_{k_0}}) + d(x_{n_{k_0}}, x*) \leq \varepsilon + \varepsilon = 2\varepsilon $$ so every Cauchy sequence converges and $K$ is complete.

Question Is this proof correct? I can't understand steps $(2)$ and $(3)$ (why uses $\frac{1}{k}$?).

Thanks in advance.

Edit Is this proof correct?

Let $\{x_n\}_{n=1}^\infty \subset K$ be a Cauchy sequence, by definition $\forall \varepsilon >0 \;\, \exists n_0\in \mathbb N \;\, \forall m,n \geq n_0 \;\, d(x_n, x_m) < \varepsilon$. Since $K$ is compact, the sequence has an accumulation point $x^*\in K$ (already proved). Let's see that $\lim_{n\to\infty}x_n = x^*$: since $x^*$ is an accumulation point, then exists a subsequence $\{x_{n_k}\}_{k=1}^{\infty}\subset \{x_{n}\}_{n=1}^{\infty}$ that converges to $x^*$, in other words $\forall \varepsilon > 0 \;\, \exists k_0\in\mathbb N \;\, \forall k \geq k_0 \;\, d(x_{n_{k}}, x^*) < \varepsilon$. Because of the triangular property $d(x_n,x^*) \leq d(x_n, x_{n_{k}}) + d(x_{n_{k}}, x^*) \leq 2\varepsilon$, so every Cauchy sequence converges and $K$ is complete.

share|cite|improve this question
What's your definition of compactness for metric spaces? There are 4 "competing" ones out there: (1) complete and totally bounded, (2) every sequence has a convergent subsequence, (3) every infinite subset has a limit point, and (4) every open cover has a finite subcover. – kahen Dec 14 '12 at 22:13
Here, definition 3 – user50554 Dec 14 '12 at 22:14
up vote 4 down vote accepted

It contains some errors, though the basic idea is right. Here’s how I would prove that $\lim\limits_{n\to\infty}x_n=x^*$.

Let $\epsilon>0$. Since the sequence is Cauchy, there is an $n_0\in\Bbb N$ such that $d(x_m,x_n)<\frac{\epsilon}2$ for all $m,n\ge n_0$. Since $x^*$ is an accumulation point of the sequence, it is also an accumulation point of the tail sequence $\langle x_n:n\ge n_0\rangle$. (Why?) Thus, there is an $n_1\ge n_0$ such that $d(x_{n_1},x^*)<\frac{\epsilon}2$. The triangle inequality then implies that $d(x_n,x^*)\le d(x_n,x_{n_1})+d(x_{n_1},x^*)<\frac{\epsilon}2+\frac{\epsilon}2=\epsilon$ for all $n\ge n_0$. Since $\epsilon>0$ was arbitrary, $\langle x_n:n\in\Bbb N\rangle\to x^*$.

share|cite|improve this answer

In (Step 2) you are essentially constructing a subsequence of $\{ x_n \}_{n=1}^\infty$ which converges to the accumulation point $x^*$ of $\{ x_n : n \in \mathbb{N} \}$. I say essentially because you are not ensuring that $n_{k+1} > n_k$, but the basic idea is there. The use of $\frac 1k$ is not overly important, but you do want some sequence of positive real numbers converging to $0$.

(Step 3) is essentially a proof of the following: Every Cauchy sequence in a metric space with a convergent subsequence is convergent (to the same limit).

share|cite|improve this answer
@user50554: I'm not quite certain if I understand your question. The main use of the $\frac 1k$ is in (Step 3) where you take some $k_0$ such that $\frac{1}{k_0} < \epsilon$ (this is implicit with $d ( x_{n_{k_0}} , x^* ) < \epsilon$). Replacing $\frac 1k$ everywhere with, say, $\frac{1}{2^k}$ would not really change the proof. – arjafi Dec 14 '12 at 22:37

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.