# Prove that every bounded sequence in $\Bbb{R}$ has a convergent subsequence [closed]

I am to prove that every bounded sequence in $\Bbb{R}$ has a convergent subsequence. I am stuck not knowing how and where to start.

-

## closed as off-topic by Carl Mummert, tired, Daniel W. Farlow, Dan, NehoraiFeb 19 at 21:03

This question appears to be off-topic. The users who voted to close gave this specific reason:

• "This question is missing context or other details: Please improve the question by providing additional context, which ideally includes your thoughts on the problem and any attempts you have made to solve it. This information helps others identify where you have difficulties and helps them write answers appropriate to your experience level." – Carl Mummert, tired, Daniel W. Farlow, Dan, Nehorai
If this question can be reworded to fit the rules in the help center, please edit the question.

16 minutes.  – Did Dec 23 '12 at 10:21

This is the so called Bolzano-Weierstrass Theorem.

I will prove that any sequence of real numbers has a monotone subsequence and leave the rest to you

First some terminology: Let $\left(a_n\right)$ be a sequence of real numbers and $m\in \mathbb{N}$. We say that $m$ is a peak of the sequence $\left(a_n\right)$ if $$n> m\implies a_n<a_m$$

Now the proof: Let $\left(a_n\right)$ be a sequence of real numbers.

Suppose that $\left(a_n\right)$ has an infinite number of peaks $$k_0<k_1<k_2<...<k_n<...$$ and consider the subsequence $(a_{k_n})$. Then, $(a_{k_n})$ is strictly decreasing since $$k_n>k_m\implies a_{k_n}<a_{k_m}$$ and thus $(a_{k_n})$ is monotone.

Suppose that $\left(a_n\right)$ has an finite number of peaks and let $N\in \mathbb{N}$ be the last (greatest) peak. Then $k_{0}=N+1$ is not a peak and so $$\exists k_1>k_0\in \mathbb{N}:a_{k_{1}}>a_{k_{0}}$$ Having defined $k_n$ such as that $k_n>k_{n-1}>N$ then $$\exists k_{n+1}>k_n\in \mathbb{N}:a_{k_{n+1}}>a_{k_{n}}$$ The subsequence $(a_{k_n})$ is obviously increasing and so it is monotone

Now if $(a_n)$ is in addition bounded, so is $(a_{k_n})$ and applying the monotone convergence theorem yields....

-
...yields that the sequence has finite limit – doniyor Dec 23 '12 at 10:13
@doniyor the subsequence – Nameless Dec 23 '12 at 10:14
yeah, the subsequence, thanks! – doniyor Dec 23 '12 at 10:15

Hint: The key result is the least-upper-bound property.

EDIT: I develop my argument. Let $(x_n)$ be a real bounded sequence and set $y_n= \sup \{ y_k : k \geq n \}$. Then $(y_n)$ is non-increasing sequence and you can show that it converges to $x=\inf \{ y_n : n \geq 0 \}$ from the definition of $\inf$. Finally, you can proved that $x$ is a limit point of $(x_n)$ and deduce that there exists a subsequence of $(x_n)$ converging to $x$.

In fact, least-upper-bound property, Bolzano-Weierstrass theorem and completness are very quite related in ordered fields.

-

Dichotomy.

Since the sequence $(x_n)_n$ is bounded, every $x_n$ is in a bounded interval $I_0$. Split $I_0$ into two subintervals of half its length. At least one of them contains infinitely many terms of the sequence, call it $I_1$. Repeat. After $n$ steps, $I_n$ is an interval of length the length of $I_0$ divided by $2^n$, and $I_n\subset I_{n-1}$. Call $I_n=(a_n,b_n)$ and pick $x_{\varphi(n)}$ in $I_n$ with $\varphi(n)\gt\varphi(n-1)$ (crucially, this is always possible since $I_n$ contains infinitely many terms of the sequence). Then $(a_n)_n$ is nondecreasing, $(b_n)_n$ is nonincreasing, $b_n-a_n\to0$ and $a_n\leqslant x_{\varphi(n)}\leqslant b_n$, hence everything is set so that the subsequence $(x_{\varphi(n)})_n$ converges.

-