# Question about arithmetic–geometric mean [duplicate]

This question already has an answer here:

We have two sequences: $$a_{n+1}=\sqrt{a_nb_n}$$ $$b_{n+1}=\frac{a_n+b_n}{2}$$ I need to prove that those are making Cantor's Lemma.(At the end I shold get that: $\lim_{n\to \infty}a_n=\lim_{n\to \infty}b_n$ by Cantor's Lemma)

Any ideas how?

Thank you.

## marked as duplicate by Arnaud D., Gibbs, José Carlos Santos calculus StackExchange.ready(function() { if (StackExchange.options.isMobile) return; $('.dupe-hammer-message-hover:not(.hover-bound)').each(function() { var$hover = $(this).addClass('hover-bound'),$msg = $hover.siblings('.dupe-hammer-message');$hover.hover( function() { $hover.showInfoMessage('', { messageElement:$msg.clone().show(), transient: false, position: { my: 'bottom left', at: 'top center', offsetTop: -7 }, dismissable: false, relativeToBody: true }); }, function() { StackExchange.helpers.removeMessages(); } ); }); }); Nov 12 '18 at 14:01

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

• Have you calculated the first few values of $a_n$ and $b_n$ for some starting values $a_1$ and $b_1$? Do you notice anything useful? – Steve Kass Nov 27 '13 at 22:22
• I tried, but I didn't know where to start... – CS1 Nov 28 '13 at 6:58

## 2 Answers

If $a_1=b_1$, the two sequences are constant, and the proof is trivial. Otherwise, suppose $a_1<b_1$ or $b_1<a_1$. We have the property that for any two numbers $a<b$, $$a<GM<AM<b,$$ where $GM$ is the geometric mean, and $AM$ is the arithmetic mean. So, we must have $$a_1<a_2<b_2<b_1$$ or $$b_1<a_2<b_2<a_1.$$ In any case, for $n>1$ $$a_n<a_{n+1}<b_{n+1}<b_n$$ That shows us that $a_n$ is increasing and $b_n$ is decreasing. To apply the Cantor's lemma we must only show that $b_n-a_n\rightarrow 0$. To show that, we realize that $b_{n+1}-a_{n+1}<\frac{b_n-a_n}{2}$, since $b_{n+1}$ is at the middle of the points $a_n$ and $b_n$.

• How do you prove it? that $a<GM<AM<b$? I don't see a way..... – CS1 Nov 28 '13 at 7:29
• I can prove it bu induction? – CS1 Nov 28 '13 at 14:43
• @YoavFridman No, you can't. Do you know what arithmetic mean is? Can you see why $\frac{a+b}{2}$ is in the middle of the interval $[a,b]$ on the real line? If yes, you should be able to prove $a\leq AM \leq b$. I explained the inequality $AM\leq GM$ under my answer. And the inequality $a\leq GM \leq b$ is obvious if you look at the squares. Alternatively, you can look at logarithms and then geometric mean becomes arithmetic mean. – savick01 Nov 28 '13 at 18:05
• As savick01 said, you can prove it with the inequality $(\sqrt b-\sqrt a)^2\geq0$, and the inequality is strict if $b>a$. – Mateus Sampaio Nov 29 '13 at 12:57
• You can prove the general inequality $AM\geq GM$ for $n$ positive numbers by induction, but is not as simple as many induction proofs. You can see it here: en.wikipedia.org/wiki/… – Mateus Sampaio Nov 29 '13 at 13:02

Hint. All you need is:

• $\sqrt{ab} \leq \frac{a+b}{2}$ -- for the inequality between the two sequences
• the distance between $\frac{a+b}{2}$ and $\sqrt{ab}$ is at most half of the distance between $a$ and $b$ because $\frac{a+b}{2}$ is just in the middle of the interval with ends $a$, $b$ and $\sqrt{ab}$ is somewhere inside it -- to show that $a_n-b_n\to 0$.
• How the first one helps me? And I don't know how to prove those two things.. If you can help me with this it will be great... – CS1 Nov 28 '13 at 7:00
• @YoavFridman The first one says that $GM\leq AM$. I skipped the fact that for $a\leq b$ we have $a\leq M\leq b$ where M is $AM$ or $GM$ as I considered it obvious. How to prove the first? Take squares of both sides of the inequality and do some calculations (you should get $(a-b)^2$ and notice that it is positive). What is unclear about the second? – savick01 Nov 28 '13 at 13:44
• I didn't understand how you prove the first one. And I didn't understand how the second one helps me... Than k you! – CS1 Nov 28 '13 at 14:41
• @YoavFridman Oh, c'mon, have you tried solving the first with the guidelines from my comment? There are like 4 steps and I gave you two of them and the other two are very simple. About the second: do you know what is the limit of $(\frac{1}{2})^n$? – savick01 Nov 28 '13 at 14:47
• I'll try. I hope I wont stuck, else I'll ask. About the second is 0, right? – CS1 Nov 28 '13 at 22:00