Is it 1 1^{\infty} ?

Calculus Level 3

If the limit lim n + ( 1 + 0 1 4 π ( 1 tan x 1 + tan x ) n d x ) n = a b \lim_{n\to\infty^+} \left(1+\int_0^{\frac{1}{4}\pi }\left(\frac{1-\tan x}{1+\tan x}\right)^ndx\right)^n = \sqrt[b]{a} where a , b a,b are positive real numbers and b b is prime. Find the value of b a b b\sqrt[b]{a} .


Inspired by Tangle .


The answer is 3.29744.

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.

1 solution

Naren Bhandari
May 5, 2020

Note that the integral 0 1 4 π ( 1 tan x 1 + tan x ) n d x = x = 1 4 π y 0 1 4 π tan n x d x \int_0^{\frac{1}{4}\pi}\left(\frac{1-\tan x}{1+\tan x} \right)^ndx\overbrace{=}^{ \text{}x=\frac{1}{4}\pi-y}\int_0^{\frac{1}{4}\pi}\tan^nxdx then substitute u = tan x d u = sec 2 x d x u=\tan x\implies du=\sec ^2 x dx and hence 0 1 4 π tan n x d x = 0 1 u n 1 + u 2 d u = k = 0 ( 1 ) k ( 0 1 u 2 k + n d u ) \int_0^{\frac{1}{4}\pi}\tan^nxdx=\int_0^1\frac{u^n}{1+u^2}du= \sum_{k=0}^{\infty}(-1)^k\left(\int_0^1 u^{2k+n}du\right) On Integration we yield k = 0 ( 1 ) k 2 k + n + 1 = k = 0 ( 1 n + 4 k + 1 1 n + 4 k + 3 ) \sum_{k=0}^{\infty}\frac{(-1)^k}{2k+n+1}=\sum_{k=0}^{\infty}\left(\frac{1}{n+4k+1}-\frac{1}{n+4k+3}\right) the latter series can be written as 1 4 k = 0 ( 1 k + 1 4 n + 4 k + 3 ) 1 4 k = 0 ( 1 k + 1 4 n + 4 k + 1 ) = 1 4 k = 0 ( 1 k + 1 1 n 1 4 + k + 1 ( 1 k + 1 1 n 3 4 + k + 1 ) ) \frac{1}{4}\sum_{k=0}^{\infty}\left(\frac{1}{k+1}-\frac{4}{n+4k+3}\right)-\frac{1}{4}\sum_{k=0}^{\infty}\left(\frac{1}{k+1}-\frac{4}{n+4k+1}\right)\\=\frac{1}{4}\sum_{k=0}^{\infty}\left(\frac{1}{k+1}-\frac{1}{\frac{n-1}{4}+k+1}-\left(\frac{1}{k+1}-\frac{1}{\frac{n-3}{4}+k+1}\right)\right) 0 1 4 π ( 1 tan x 1 + tan x ) n d x = 1 4 ( H n 1 4 H n 3 4 ) \therefore \int_0^{\frac{1}{4}\pi}\left(\frac{1-\tan x}{1+\tan x}\right)^ndx=\frac{1}{4}\left(H_{\frac{n-1}{4}}-H_{\frac{n-3}{4}}\right) Using the fact ( the asymptotic expansion) of H n γ + ln n + 1 2 n O ( n 2 ) H_n\approx \gamma +\ln n+\frac{1}{2n}-O(n^{-2}) we yield 1 4 ( H n 1 4 H n 3 4 ) 1 2 n 0 as n \frac{1}{4}\left(H_{\frac{n-1}{4}}-H_{\frac{n-3}{4}}\right)\approx \frac{1}{2n}\to 0 \;\text{as}\; n\to \infty and hence the required limit is lim n ( 1 + 1 2 n ) n = e \sim\lim_{n\to \infty}\left(1+\frac{1}{2n}\right)^n=\sqrt{e} and hence b a b = 2 e 3.219744 b\sqrt[b]{a}=2\sqrt{e}\approx 3.219744 .

A nice question......even I had generalised the integral after solving Tangle, but never thought of posting a question...!!

Aaghaz Mahajan - 1 year, 1 month ago

Log in to reply

That's good to know. You can even advanced this problem if you wish. 😉

Naren Bhandari - 1 year, 1 month ago

[ Redacted message, because I made a mistake ]

Pi Han Goh - 1 year, 1 month ago

Log in to reply

How did you reach 1 2 n \frac{1}{2n} by solving the series?

Aaghaz Mahajan - 1 year, 1 month ago

Log in to reply

1 n 1 1 n 3 1 2 n \frac1{n-1} - \frac1{n-3} \to \frac1{2n}

Pi Han Goh - 1 year, 1 month ago

Log in to reply

@Pi Han Goh Hey wait, this doesn't hold. Lemme work it out....

Pi Han Goh - 1 year, 1 month ago

Log in to reply

@Pi Han Goh Yeah that's what i mentioned in my previous comment but then deleted it....

Aaghaz Mahajan - 1 year, 1 month ago

Log in to reply

@Aaghaz Mahajan Sorry, I made a booboo. I assumed n n is an integer, which is a fallacy. I've retracted my message.

Pi Han Goh - 1 year, 1 month ago

Log in to reply

@Pi Han Goh So, the harmonic number approach is correct? I just want to confirm, because I reached the same value of the integral, as Naren did....

Aaghaz Mahajan - 1 year, 1 month ago

Log in to reply

@Aaghaz Mahajan Yup correct. I would have solved it via digamma functions though.

Pi Han Goh - 1 year, 1 month ago

Well, my ideas was to generalize the integral so I did above. I agree we dont need to play around harmonic numbers or digamma function if we just care about the limit to be solved. As I alternative solution, I solved using DCT, by squeeze theorem using Bernoulli's inequality :)

Naren Bhandari - 1 year, 1 month ago

Log in to reply

NICEEEEEEEEEEEE

Pi Han Goh - 1 year, 1 month ago

Log in to reply

@Pi Han Goh Thank you sir. I have written a note ( outline of general problem) here . Please have a look at it. :)

Naren Bhandari - 1 year, 1 month ago

Log in to reply

@Naren Bhandari Great work.... I'm still digesting it....

Pi Han Goh - 1 year, 1 month ago

I solved the problem like this way:

Let I n = 0 π 4 tan n x I_n=\int_{0}^{\frac{\pi}{4}} \tan^nx , therefore we get the the relation I n + I n 2 = 1 n 1 I_n+I_{n-2}=\dfrac{1}{n-1} . Now, since n n \to \infty , I n I n 2 I_n \approx I_{n-2} . This implies I n 1 2 ( n 1 ) I_n \approx \dfrac{1}{2(n-1)} .

Now from here, by same approach we can reach to the answer of e \sqrt{e} .

@Pi Han Goh @Aaghaz Mahajan

Vilakshan Gupta - 1 year, 1 month ago

Log in to reply

That's clever idea.

Naren Bhandari - 1 year, 1 month ago

Hahahaha, how did I missed this?!

Thanks!

Pi Han Goh - 1 year, 1 month ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...