Does the note matter?

Calculus Level 2

Consider the sequence a 1 , a 2 , a_1,a_2,\ldots where a n = cos ( π n 2 + n ) a_n = \cos(π\sqrt{n^{2}+n}) . Find lim n a n . \large \lim_{n\rightarrow\infty} a_n.

0 -1 1 Does not exist

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

4 solutions

Note first that

n 2 + n = ( n + 1 2 ) 2 1 4 = 2 n + 1 2 1 1 ( 2 n + 1 ) 2 = \sqrt{n^{2} + n} = \sqrt{(n + \dfrac{1}{2})^{2} - \dfrac{1}{4}} = \dfrac{2n + 1}{2}\sqrt{1 - \dfrac{1}{(2n + 1)^{2}}} =

( n + 1 2 ) 1 1 ( 2 n + 1 ) 2 . \left(n + \dfrac{1}{2}\right)\sqrt{1 - \dfrac{1}{(2n + 1)^{2}}}.

In the limit as n n \rightarrow \infty this expression goes to ( n + 1 2 ) , \left(n + \dfrac{1}{2}\right), and so

lim n cos ( π n 2 + n ) = cos ( π ( lim n n 2 + n ) ) = \lim_{n \rightarrow \infty} \cos(\pi\sqrt{n^{2} + n}) = \cos(\pi * (\lim_{n \rightarrow \infty} \sqrt{n^{2} + n})) =

lim n cos ( π ( n + 1 2 ) ) = 0 \lim_{n \rightarrow \infty} \cos(\pi*\left(n + \dfrac{1}{2}\right)) = \boxed{0}

since for any integer n n we have cos ( π ( n + 1 2 ) ) = 0. \cos(\pi\left(n + \dfrac{1}{2}\right)) = 0.

Moderator note:

Note that it is not always true that

lim n f ( g ( n ) ) = f ( lim n g ( n ) ) \lim_{ n \rightarrow \infty} f(g(n)) = f\left( \lim_{ n \rightarrow \infty} g(n) \, \right)

More justification needs to be provided.

Nice solution :)

Kïñshük Sïñgh - 6 years, 1 month ago

Log in to reply

Thanks. Nice problem: I almost quickly chose "does not exist" before I gave it a second thought. :)

Brian Charlesworth - 6 years, 1 month ago

Log in to reply

I have 1 step solution for this :)

Krishna Sharma - 6 years, 1 month ago

Yes sir, this problem is tricky :D... Many would choose limit doesn't exists.

Kïñshük Sïñgh - 6 years, 1 month ago

This might be a really stupid question, but why is lim n n 2 + n n \lim \limits_{n \to \infty} \sqrt{n ^ 2 + n} \neq n as n 2 + n = n 1 + 1 n \sqrt{n ^ 2 + n} = n * \sqrt{1 + \frac {1}{n}} and ( 1 + 1 n ) 0 if n . (1 + \frac {1}{n}) \rightarrow 0 \: \text {if} \: n \rightarrow \infty .

What am I missing here?

Ameya Daigavane - 5 years, 11 months ago

Log in to reply

The Taylor series for ( 1 + x ) α (1 + x)^{\alpha} is 1 + α x + O ( x 2 ) , 1 + \alpha x + O(x^{2}), so

1 + 1 n = ( 1 + 1 n ) 1 2 = 1 + 1 2 n + O ( 1 n 2 ) . \sqrt{1 + \frac{1}{n}} = (1 + \frac{1}{n})^{\frac{1}{2}} = 1 + \frac{1}{2n} + O(\frac{1}{n^{2}}).

When we multiply this by n n we have an expression of the form

n + 1 2 + O ( 1 n ) , n + \frac{1}{2} + O(\frac{1}{n}),

which goes to n + 1 2 n + \frac{1}{2} as n . n \rightarrow \infty.

Brian Charlesworth - 5 years, 11 months ago

Log in to reply

I think I got it. The idea is to not evaluate the limit, but express it in terms of n . n.

Ameya Daigavane - 5 years, 11 months ago

Very nice!

andy taylor - 3 years, 6 months ago

Nice one. :) Now consider this:

n 2 + n = n 1 + 1 n . \sqrt{n^2 +n} = n\sqrt{1 + \frac{1}{n}}.

By the same horribly faulty reasoning you obtain that the limit in question doesn't exist.

Vít Tuček - 5 years, 2 months ago

but why can't we consider this as n->inf , n^2>>>n hence cos(pi*root(n^2+n)= cos(npi)=hence, +/- 1 !

Arjun SivaÞrasadam - 5 years, 5 months ago

Log in to reply

It is tempting to ignore the n n term as n n \rightarrow \infty , but it does make a big difference when we take the square root and then multiply by π \pi . To look at a concrete example, consider n = 1000 n = 1000 . If we ignore the n n term then we have 100 0 2 = 1000 \sqrt{1000^{2}} = 1000 , and cos ( 1000 π ) = 1 \cos(1000*\pi) = 1 . But including the n n term gives us 100 0 2 + 1000 = 1000.499875.... \sqrt{1000^{2} + 1000} = 1000.499875.... , and cos ( 1000.499875 π ) = 0.0003927 , , , \cos(1000.499875\pi) = 0.0003927,,,

Since n 2 + n = ( n + 1 2 ) 2 1 4 n^{2} + n = \left(n + \dfrac{1}{2}\right)^{2} - \dfrac{1}{4} , we have that n 2 + n \sqrt{n^{2} + n} goes to n + 1 2 n + \dfrac{1}{2} as n n \rightarrow \infty , and this extra 1 2 \dfrac{1}{2} term makes all the difference when multiplying by π \pi .

Brian Charlesworth - 5 years, 5 months ago

Log in to reply

Thanks I think i got the concept . But i still remember getting all similar limit correctly by this method even in my university class. But i think the trig function make the whole difference !! Thanks for Siting an example, that was much self explanatory !! Brian

Arjun SivaÞrasadam - 5 years, 4 months ago

Log in to reply

@Arjun SivaÞrasadam You're welcome. :)

Brian Charlesworth - 5 years, 4 months ago

We can do it .. also By this way Y = cos (π√n² + n) dy/dn = –sin(π√n² + n)× [(π/√n² + n)×2n + 1 Now put n= ∞ You get ∞ × 0 = [0] Ans.

Ravi Hammad - 5 years, 2 months ago

Log in to reply

Wow...no words...

andy taylor - 3 years, 6 months ago

Note that it is not always true that

lim n f ( g ( n ) ) = f ( lim n g ( n ) ) \lim_{ n \rightarrow \infty} f(g(n)) = f\left( \lim_{ n \rightarrow \infty} g(n) \, \right)

More justification needs to be provided.

Calvin Lin Staff - 4 years, 10 months ago

Isn't this true if n is natural? if n is a real number then pi*(n+0.5) is not necessary a value in which cos function is zero.

Vlad Tucaliuc - 2 years, 11 months ago

Log in to reply

Yes. Note that the series a n a_n is indexed by integers, and so the limit is taken over the integers.

Calvin Lin Staff - 2 years, 11 months ago

Log in to reply

Yes, I was blind. Now I see that the oblique asymptote at +inf is n+0.5. It was easy but if your not careful you say it DOES NOT EXIST.

Vlad Tucaliuc - 2 years, 11 months ago
Krishna Sharma
Apr 24, 2015

Ok,so here is the shorter solution

First, take n 2 n^2 common from the root

lim n c o s ( π n ( 1 + 1 n ) 1 2 ) \displaystyle \lim_{n \to \infty } cos(\pi n \left( 1 + \dfrac{1}{n} \right)^{\frac{1}{2}})

Now, I use this property the most

( 1 + x ) n = 1 + n x , x 1 (1+ x)^n = 1 + nx, x \ll 1

Here

lim n c o s ( π n ( 1 + 1 2 n ) ) = lim n s i n ( n π ) \displaystyle \lim_{n \to \infty } cos( \pi n \left( 1 + \dfrac{1}{2n} \right) ) = \lim_{n \to \infty } sin( n\pi)

Since n is an integer lim n s i n ( n π ) = 0 \displaystyle \lim_{n \to \infty } sin( n \pi) = 0 .

Short but sweet. :)

Brian Charlesworth - 6 years, 1 month ago

A good solution but my way was a bit irrational or wrong too.(Cause I do not have a deep knowledge on limits)

A Former Brilliant Member - 6 years, 1 month ago

That's Bernoulli's Inequality

Radinoiu Damian - 6 years ago

Log in to reply

What is Bernoulli's inequality?

Puneet Pinku - 4 years ago
Carsten Meyer
Feb 19, 2019

This proof does not rely on unexplained exchanges of limites. First examine the argument:

π n 2 + n = π ( n 2 + n n ) + n π = 3. Binomi π n 2 + n n 2 n 2 + n + n + n π = π 1 1 + n 2 + 1 + n π first part π 2 = π ( 1 1 + n 2 + 1 1 2 ) + ( n + 1 2 ) π = : c n + 2 n + 1 2 π , c n 0 for n \begin{aligned} \pi\sqrt{n^2+n} &= \pi(\sqrt{n^2+n}-n)+n\pi\underset{\text{3. Binomi}}{=}\pi\frac{n^2+n - n^2}{\sqrt{n^2+n}+n}+n\pi=\pi\frac{1}{\sqrt{1+n^{-2}}+1}+n\pi\qquad\left| \text{first part }\rightarrow\:\frac{\pi}{2} \right.\\\\ &=\pi\left( \frac{1}{\sqrt{1+n^{-2}}+1}-\frac{1}{2} \right)+\left(n+\frac{1}{2}\right)\pi=:c_n+\frac{2n+1}{2}\pi,\qquad |c_n|\rightarrow 0\quad\text{for }n\rightarrow\infty \end{aligned}

Use that to gain an upper estimate of the given sequence:

0 a n = cos ( 2 n + 1 2 π + c n ) = ( 1 ) n cos ( π 2 + c n ) = sin ( c n ) = sin c n 0 a n 0 \begin{aligned} 0\leq |a_n|&=\left| \cos\left(\frac{2n+1}{2}\pi+c_n\right) \right|=\left| (-1)^n\cos\left(\frac{\pi}{2}+c_n\right) \right|=|-\sin(c_n)|=\sin|c_n|\rightarrow 0\\\\ \Rightarrow \quad a_n&\rightarrow\fbox{0} \end{aligned}

Rem.: In the last step, we use that sin ( x ) \sin(x) is continuous in x = 0 x=0 so that we can be sure the following limit exists:

lim x 0 sin ( x ) = 0 \begin{aligned}\lim_{x\rightarrow 0}\sin(x)=0\end{aligned}

Hunter Seropian
Jul 20, 2016

Believe it or not I just used the squeeze theorem on this problem, and i got it right...

Moderator note:

Without knowing exactly what you did, I am unable to comment about the validity of your solution.

Without knowing exactly what you did, I am unable to comment about the validity of your solution. There are many ways to use the squeeze theorem, and some of them need not lead to a completely rigorous solution.

Calvin Lin Staff - 4 years, 10 months ago

At first attempt, I got it doesn't exist. Then, unsatisfied by the comments, I went to Wolfram Alfa and tried to computer the limit. The computer said it lies between -1 and 1. It's true right. I also checked other limit calculators and they gave same answer. May be the answer is wrong.

Sai Srikar V - 2 years, 5 months ago

Log in to reply

Did you take the limit over the reals, or over the integers?

If you took it over the reals, then the limit does not exist, since we "essentially" still have a cosine graph.

Calvin Lin Staff - 2 years, 5 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...