In a few hours I will have a quiz and while studying I had some questions. Could you please help me?

Thanks in advance

Question 1: Let $(X, d)$ be a metric space and $(x_n)$ be a sequence in $X$. Let $(x_n)$ be a Cauchy sequence, prove that if $(x_n)$ has a cluster point $z$, then $(x_n) \to z$.

My attempt: Suppose that $z$ is a cluster point of $(x_n)$, and construct a subsequence $x_{n_k}$ with limit $z$.

Choose $x_{n_1} \in B(z,1)$, $B(z,1/2)$ contains infinitely many elements of $(x_n)$

Choose $x_{n_2} \in B(z,1/2)$ such that $n_2 > n_1$

..... ..... .....

Choose $x_{n_k} \in B(z,1/k)$ so that $n_k > n_{k-1} > \cdots$

By the choice we did, $d(x_{n_k},z) < 1/k$ and we know that $1/k \to 0$ then, $d(x_{n_k},z) \to 0$. Hence, $x_{n_k} \to z$.

Now how can I show that $x_n$ also converges to $z$?

Question 2: Let $a_n > 0$, $b_n > 0$, $n=1,2,\dots$. Prove that if $\frac{a_n}{b_n} \to l \neq 0$, then the series $\sum_{n=1}^{\infty} a_n$ and $\sum _{n=1}^{\infty} b_n$ converge or diverge simultaneously.

My attempt: If $\frac{a_n}{b_n} \to l \neq 0$ then we can also say that if $\frac{a_n}{b_n}$ is closed by $l$.

If $l>0$, then there are positive numbers $c$ and $d$ with $0<c\leq l \leq d$ with $c \leq \frac{a_n}{b_n} \leq d$. For $b_n$>0, $cb_n \leq a_n \leq db_n$.

Now if $\sum_{n=1}^{\infty} b_n$ converges, then so does $\sum_{n=1}^{\infty} db_n$. From the inequality $\frac{a_n}{b_n} \leq d$, $\sum_{n=1}^{\infty} a_n$ also converges. However if $\sum_{n=1}^{\infty} b_n$ diverges, then so does $\sum_{n=1}^{\infty } cb_n$.

From $a_n \leq db_n$ we see that if $\sum_{n=1}^{\infty} a_n$ diverges then $\sum_{n=1}^{\infty} \frac{a_n}{d}$ also diverges and in this case $\sum_{n=1}^{\infty} b_n$ diverges too. With similar argument we can see that if $\sum_{n=1}^{\infty} a_n$ converges then $\sum_{n=1}^{\infty} \frac{a_n }{c}$ also converges and from the left inequality ($b_n < \frac{a_n}{c}$) $\sum_{n=1}^{\infty} b_n$ converges too.

Is this proof right and enough?

  • 5
    $\begingroup$ These are two entirely different questions in different topics. Please delete one question (edit) and re-post that question as a separate question. $\endgroup$ – Namaste Nov 26 '12 at 0:35
  • 1
    $\begingroup$ Excuse me, that's THREE separate questions in two different topics, in ONE post? $\endgroup$ – Namaste Nov 26 '12 at 1:00
  • $\begingroup$ @amWhy you are right. I was a little bit panicked to get a fast answer that is why I did such a thing. I have already got answers for the first two questions. So I only deleted the third question and will post it as a separate question. $\endgroup$ – Amadeus Bachmann Nov 26 '12 at 1:17
  • $\begingroup$ "if $\frac{a_n}{b_n}$ is closed by $l$" (as in line 1 of your attempt on question 2) does not mean anything, and it is not even a statement. Maybe you meant to say that if $a_n/b_n\to l$ then $a_n/b_n$ is close to $l$? The problem is that even though this informal statement is true (for large $n$, anyway), it is useless, precisely because it is informal. $\endgroup$ – Andrés E. Caicedo Nov 26 '12 at 1:49
  • $\begingroup$ @AndresCaicedo yes I meant to say "it is close to $l$" but now I am aware that that statement was useless. $\endgroup$ – Amadeus Bachmann Nov 26 '12 at 2:00

For the series question use the definition of the limit of a sequence. Let $\epsilon=\frac{l}{2}$, then

$$ \left| \frac{a_n}{b_n} - l\right| < \frac{l}{2}\quad \mathrm{when}\> n>N $$

$$ \implies \frac{l}{2}<\frac{a_n}{b_n}<\frac{3l}{2} \quad \mathrm{when}\> n>N $$

$$ \implies \frac{l}{2} b_n< a_n <\frac{3l}{2}b_n \quad \mathrm{when}\> n>N $$

Now, if $\sum b_n $ converges, so does $\sum \frac{3l}{2}b_n $. Then the right half of the above inequality shows that $\sum_{n=N}^{\infty}a_n$ converges by the comparison test. It follows that $\sum_{1}^{\infty}a_n $ converges. I think you can finish the proof now.

  • $\begingroup$ @Zxy: You are welcome. $\endgroup$ – Mhenni Benghorbal Nov 26 '12 at 1:14
  • $\begingroup$ @MehenniBenghorbal Oh it just came to my mind: is it OK to directly say that let $\epsilon=\frac{l}{2}$ for a formal proof? I know it does not change anything as long as $\epsilon$>0. But I had a feeling that my professor would not be happy if I directly say let $\epsilon=\frac{l}{2}$. = ) $\endgroup$ – Amadeus Bachmann Nov 26 '12 at 2:08

I'll answer the first question. You have a Cauchy sequence that has a convergent subsequence, and you'd like to show the sequence converges. So, let $\epsilon>0$. Since $(x_n)$ is Cauchy, we can choose $N_1$ so that for every $n,m\geq N_1$ we have $d(x_n,x_m)<\frac{\epsilon}{2}$. Then, since the subsequence $(x_{nk})$ converges to $z$, we can choose $N_2$ so that for every $k\geq N_2$, $d(x_{nk},z)<\frac{\epsilon}{2}$. Then, set $N=\max\{N_1,N_2\}$ so that for every $n>N$, we have by the triangle inequality

$$ d(x_n,z)\leq d(x_n,x_{nn})+d(x_{nn},z)<\frac{\epsilon}{2}+\frac{\epsilon}{2}=\epsilon $$

Thus $(x_n)$ indeed converges to $z$.

  • $\begingroup$ I am very grateful for your help! $\endgroup$ – Amadeus Bachmann Nov 26 '12 at 1:14

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.