3
$\begingroup$

I'm not quite sure how to tackle these kinds of questions in general, but I tried something that I thought could be right. Hoping to be steered in the right direction here!

Let $\{\frac{3n+5}{2n+6}\}$ be a sequence of real numbers. Prove that this sequence is Cauchy.

Proof:

We want to establish that $\forall_{\epsilon>0}\exists_{{n_0}\in{\mathbb{N}}}\forall_{n,m\geq n_0}\big(|f(n)-f(m)|\big)<\epsilon.$

Suppose $n>m$ without loss of generality. We then know that $\frac{3n+5}{2n+6}>\frac{3m+5}{2m+6}$ and thus that $\frac{3n+5}{2n+6}-\frac{3m+5}{2m+6}>0$ such that $\frac{3n+5}{2n+6}-\frac{3m+5}{2m+6}=|\frac{3n+5}{2n+6}-\frac{3m+5}{2m+6}|=|f(n)-f(m)|.$

Let us work out the original sequence:

$\frac{3n+5}{2n+6}-\frac{3m+5}{2m+6}=\frac{(3n+5)(2m+6)-(3m+5)(2n+6)}{(2n+6)(2m+6)} = \frac{8(n-m)}{(n2+6)(2m+6)}<\frac{8(n-m)}{nm}= 8(\frac{1}{n}- \frac{1}{m}).$

We know that $\frac{1}{n}<\frac{1}{m}$ as $n>m$ and that $\frac{1}{n}\leq\frac{1}{n_0}, \frac{1}{m}\leq\frac{1}{n_0}$ for $n,m\geq n_0$.

This means that $\frac{1}{n}-\frac{1}{m}\leq \frac{1}{n_0}- \frac{1}{m}\leq\frac{1}{n_0}$, and thus $8(\frac{1}{n}- \frac{1}{m})\leq \frac{8}{n_0}$.

Let $\epsilon=\frac{8}{n_0}$, as it only depends on $n_0$ it can become arbitrarily small.

Then the following inequality holds: $\frac{3n+5}{2n+6}-\frac{3m+5}{2m+6}<8(\frac{1}{n}- \frac{1}{m})\leq \frac{8}{n_0}$.

So: $\frac{3n+5}{2n+6}-\frac{3m+5}{2m+6}<\epsilon$, and thus the sequence is Cauchy.$\tag*{$\Box$}$

$\endgroup$
  • $\begingroup$ Any convergent sequence is Cauchy. $\endgroup$ – Bernard Jan 25 '18 at 11:49
  • 2
    $\begingroup$ Your proof isn't correct. $\epsilon$ must be arbitrary and specified in advance. Your choice of $n_0$ should depend only on $\epsilon$. $\endgroup$ – Umberto P. Jan 25 '18 at 11:50
  • $\begingroup$ What if you don't know if the sequence is convergent? Besides, doesn't that only apply to real numbers (so not for example only rational numbers?) $\endgroup$ – Marc Jan 25 '18 at 11:51
  • 1
    $\begingroup$ @Marc that's the other way around. Convergent always implies Cauchy in any metric space. $\endgroup$ – Umberto P. Jan 25 '18 at 11:52
  • $\begingroup$ @Marc Every Cauchy sequence of rational numbers converges to a real number. $\endgroup$ – José Carlos Santos Jan 25 '18 at 11:52
2
$\begingroup$

$|\frac{3n+5}{2n+6}-\frac{3m+5}{2m+6}|=\frac{|n-m|}{(2n+6)(2m+6)}\le\frac{8|n-m|}{nm} =8|\frac{1}{n}-\frac{1}{m}|$.

Can you proceed ?

$\endgroup$
  • $\begingroup$ Thank you! I was almost there before, trying to work out some algebraic magic, but I messed up the m's and n's. I edited my post, is it now correct? $\endgroup$ – Marc Jan 25 '18 at 12:36
2
$\begingroup$

If you do have to prove it with the ε-$n_0$ method, it's much simpler that that:

First rewrite $a_n$ as $$a_n=\frac{3n+5}{2n+6}=\frac32-\frac2{n+3}.$$ Then, if $m,n>n_0$, $$\bigl|a_m-a_n\bigr|=2\,\biggl|\frac1{m+3}-\frac1{n+3}\biggr|=\frac{ 2\,| m-n|}{(m+3)(n+3)}\le 2\biggl|\frac1m-\frac1n\biggr|<2\biggl(\frac1m+\frac1n\biggr)<\frac4{n_0},$$ and this one is less than $\varepsilon$ as soon as $\;n_0>\dfrac 4\varepsilon$.

$\endgroup$
  • $\begingroup$ @lhf: You're right. I'll fix that in a moment. $\endgroup$ – Bernard Jan 25 '18 at 12:15
2
$\begingroup$

As you were told in the comments, you must take some $\varepsilon>0$ and then prove that $\left|\frac{3n+5}{2m+6}-\frac{3n+5}{2n+6}\right|<\varepsilon$ if $m$ and $n$ are large enough.

Note that$$(\forall n\in\mathbb{N}):\frac{3n+5}{2n+6}=\frac{3n+9}{2n+6}-\frac4{2n+6}=\frac32-\frac2{n+3}.$$So, take $p\in\mathbb N$ such that $\frac2{p+3}<\frac\varepsilon2$. Then$$m,n\geqslant p\implies\left|\frac{3m+5}{2m+6}-\frac{3n+5}{2n+6}\right|=\left|\left(\frac32-\frac2{m+3}\right)-\left(\frac32-\frac2{n+3}\right)\right|<\varepsilon.$$

$\endgroup$
1
$\begingroup$

Your proof doesn't use any properties of the sequence: you saying that the sequence is Cauchy is no proof. Also, your first sentence is wrong, as the sequence is increasing.

What you want to do is $$ \frac {3n+5}{2n+6}-\frac {3m+5}{2m+6}=\frac{8 (n-m)}{(2m+6)(2n+6)} \leq \frac{8 (n-m)}{4mn}=\frac2m-\frac2n. $$ Now you can show that, the bigger $n,m $ are, the smaller the difference. Usually one needs absolute value, but in this particular case the condition $n>m $ guarantees that everything above is positive.

$\endgroup$
0
$\begingroup$

Let $ \epsilon >0$ be an arbitrary small number.

Choose a positive integer $n_0$ such that $$\frac {2}{n_0 +3} <\frac {\epsilon }{2}$$

Note that if $m \ge n_0$, and $n\ge n_0$ are integers, then $$ |f(n)-f(m)| = |\frac{3n+5}{2n+6}-\frac{3m+5}{2m+6}|= $$ $$|\frac {2(m-n)}{(n+3)(m+3)} |\le $$ $$\frac {2(m+n)}{(n+3)(m+3)} <$$ $$\frac {2}{(m+3)}+\frac {2}{(n+3)} <\epsilon $$

Thus the sequence ($ \frac{3n+5}{2n+6}$) is a Cauchy sequence.

$\endgroup$

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.