1
$\begingroup$

Show that if $x_n \le y_n \le z_n$ for all $n \in \mathbb N$, and if $lim(x_n) = lim (z_n) = l$, then $lim (y_n)= l$ as well.

Proof. Since $lim(x_n) = lim (z_n) = l$. Then follows that $|x_n -l| < \frac{\epsilon}{2}$ whenever $n \ge N_1$ and similarly $|l-z_n | < \frac{\epsilon}{2}$ whenever $n \ge N_2$. Choose $max \{ N_1,N_2\}$ and by the triangle inequality we get $$|x_n - z_n| \le|x_n -l|+ |l-z_n| \lt \epsilon .$$

since $$|x_n - z_n| \lt \epsilon $$ it follows that $x_n=z_n$ for sufficiently large $n \ge \max\{N_1, N_2\}$. Hence it follows that $x_n=y_n= z_n$ (from the fact that $x_n=z_n$ and $x_n \le y_n \le z_n$). And therefore $lim(x_n) = lim (y_n) = l$.

Question 1: Is my attempt at the proof correct? If not why not? Also if not can you give a correct version or atleast a few hints?

Question 2: I use the property that $|a-b|< \epsilon$ is equivalent to $a=b$ but it doesn't sit well with me below that $x_n=z_n$ since $(x_n), (z_n)$ could have different elements and still converge at the same value.) Why am I getting contradictory results here?

Thanks in advance.

I really appreciate this site helping me learn how to write proofs in real analysis :)

$\endgroup$
1
$\begingroup$

Firstly, $z_n$ does not equal $x_n$ for sufficiently large $n$. For an answer to your second question $\frac{1}{n}$ and $\frac{1}{2n}$ both converge to zero, but they are never equal for a given $n$. Use the definition of convergence instead. What do you need to show that $\lim_{n\to \infty} y_n=l$?

$\endgroup$
  • $\begingroup$ Yes that is correct $\endgroup$ – Red Jul 15 '18 at 2:55
  • $\begingroup$ I completely agree with you that $z_n$ does not necessarily equal to $x_n$ but as I have shown the inequality $|x_n - z_n| \lt \epsilon$ would imply that $x_n = z_n$ doesn't it?? $\endgroup$ – Red Jul 15 '18 at 2:59
  • $\begingroup$ @Red: why would the inequality imply $z_n=x_n$? $\endgroup$ – Paramanand Singh Jul 15 '18 at 6:06
1
$\begingroup$

Most elementary results of convergence for sequences either depend on the triangle inequality (via an $\epsilon/2$ trick) or the inequality $0\leq\frac{|x|}{1+|x|}\leq 1$ for all $x$ (to avoid cases where division by $0$ may occur).

The Squeeze Theorem however is probably best proved by a different approach.

Try to show that if $x\leq y\leq z$ then $|y-l|\leq \max\left\{|x-l|, |z-l|\right\}$ no matter what $l$ may be. This is probably best proved with a picture.

$\endgroup$
  • $\begingroup$ Is the aproach I have taken completely wrong or can it be tweaked for a correct proof? $\endgroup$ – Red Jul 15 '18 at 3:11
  • $\begingroup$ @Red you have claimed that the sequences $\{x_n\}$ and $\{z_n\}$ eventually agree. But an easy counter example is $x_n=-1/n$, $z_n=1/n$ and $y_n=0$ for all $n$. $\endgroup$ – Robert Wolfe Jul 15 '18 at 3:12
  • $\begingroup$ I completely agree with this, the sequences need not agree eventually.But why is the triangle inequality gving me that result anyway? I feel like I have used it wrong but I cannot figure out where my mistake llies? as I said in th second part of my question $\endgroup$ – Red Jul 15 '18 at 3:16
  • 1
    $\begingroup$ @Red if $a$ and $b$ satisfy $|a-b|<\epsilon$ for every $\epsilon>0$ then $a=b$. However you have not proved that $|x_n-z_n|<\epsilon$ for every $\epsilon>0$. You have proved that for every $\epsilon>0$ there is an $n$ such that $|x_n-z_n|<\epsilon$. $\endgroup$ – Robert Wolfe Jul 15 '18 at 3:27
  • $\begingroup$ I see my mistake. Thanks! $\endgroup$ – Red Jul 15 '18 at 3:28

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.