The Limit Of The Exponent

Calculus Level 3

It is well known that the sequence ( 1 + 1 n ) n ( 1 + \frac{1}{n} ) ^ n approaches e e from below, and the sequence ( 1 + 1 n ) n + 1 ( 1 + \frac{1}{n} ) ^ {n+1} approaches e e from above. Hence, for each n n , there is a unique value a n a_n between 0 and 1 such that ( 1 + 1 n ) n + a n = e ( 1 + \frac{1}{n}) ^ {n+ a_n} = e .

Determine lim n a n \displaystyle\lim_{n \rightarrow \infty} a_n .

1 0 Does not exist 1 2 \frac{1}{2}

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

Leon Fernandes
Feb 21, 2014

We have been given that ( 1 + 1 n ) n + a n = e ( 1+ \frac{1}{n})^{n+a_n} = e .

My first attempt is to solve for for a n a_{n} .

Taking l n ln on both sides,

( n + a n ) l n ( 1 + 1 n ) = 1 (n + a_n) * ln ( 1 +\frac{1}{n} ) = 1

So, ( n + a n ) = 1 l n ( 1 + 1 n ) (n+ a_n) = \frac{1}{ln(1+\frac{1}{n})}

Thus , a n = 1 l n ( 1 + 1 n ) n a_n = \frac{1}{ ln (1+\frac{1}{n})} - n

We have to find lim n a n = ? \lim_{n \to \infty} \ a_n = ?

i.e we will have to find

lim n 1 l n ( 1 + 1 n ) n \lim_{n \to \infty} \frac{1}{ ln (1 + \frac{1}{n})} - n

We can treat it as a whole new problem. I prefer to change variable from n n \to \infty to x 0 x \to 0 by putting x = 1 n x = \frac{1}{n}

lim x 0 + 1 l n ( 1 + x ) 1 x \lim_{x \to 0^{+} } \frac{1}{ln(1+x)} - \frac{1}{x}

= lim x 0 + x l n ( 1 + x ) x l n ( 1 + x ) = \lim_{x \to 0^{+} } \frac{x - ln(1+x)}{x*ln(1+x)}

Applying l'Hopital's rule ( i.e. differentiate the numerator and denominator w.r.t x ) as the expression is of the form 0 0 \frac{0}{0} ,

lim x 0 + 1 1 1 + x l n ( 1 + x ) + x 1 + x \lim_{x \to 0^{+} } \frac{1 - \frac{1}{1+x}}{ln(1+x) + \frac{x}{1+x}}

= lim x 0 + 1 + x 1 ( 1 + x ) ( l n ( 1 + x ) + x = \lim_{x \to 0^{+} } \frac{1+x-1}{(1+x)*(ln(1+x) +x }

= lim x 0 + x ( 1 + x ) ( l n ( 1 + x ) + x = \lim_{x \to 0^{+} } \frac{x}{(1+x)*(ln(1+x) +x }

Again we see that at x = 0 x = 0 the expression is indeterminate ( 0 0 ) (\frac{0}{0}) .

On second application of l'Hopital's rule ,

lim x 0 + 1 1 1 + x + l n ( 1 + x ) + x 1 + x + 1 \lim_{x \to 0^{+} } \frac{1}{\frac{1}{1+x} + ln ( 1+x ) + \frac{x}{1+x} + 1}

This expression at x = 0 gives value : 1 1 + 0 + 0 + 1 = 1 2 . \frac{1}{1+0+0+1} = \frac{1}{2}.

Therefore we get lim n a n = \lim_{n \to \infty} \ a_n = 1 2 \boxed{\frac{1}{2}}

well-written!

Victor Loh - 7 years, 3 months ago

Nice!

Finn Hulse - 7 years, 3 months ago

good

Sharma Manikanta - 7 years, 3 months ago

excelent.........

Alamin Sheikh - 7 years, 3 months ago

thats great !

sanket doshi - 7 years, 3 months ago

Neat solution

Usha Ramakrishnan - 7 years, 3 months ago

very simple and neat solution

Anirudh Kasuganti - 7 years, 3 months ago

very simple soln

Amer Iliyas - 7 years, 3 months ago

brilliant :D

Christian Baldo - 7 years, 1 month ago

So simple

Arshad Kamal - 7 years, 1 month ago
Oliver Bel
Mar 30, 2014

Rearranging ( 1 + 1 n ) n + a n = e \left(1+\frac{1}{n}\right)^{n+a_n}=e and taking limits gives lim n a n = lim n ( 1 n log ( 1 + 1 n ) log ( 1 + 1 n ) ) \lim_{n\to\infty}a_n=\lim_{n\to\infty}\left(\frac{1-n\log{\left(1+\frac{1}{n}\right)}}{\log\left(1+\frac{1}{n}\right)}\right)

Using the Maclaurin series log ( 1 + x ) = r = 1 ( 1 ) r + 1 x r r \displaystyle\log(1+x)=\sum_{r\mathop=1}^{\infty}\frac{(-1)^{r+1}x^r}{r} it is easy to see the limit tends to 1 2 \boxed{\dfrac{1}{2}}

Tim Ratigan
Feb 24, 2014

We have lim n ( 1 + 1 n ) n + a n = e \lim_{n\to\infty}\left(1+\frac{1}{n}\right)^{n+a_n}=e

Taking the log \log of both sides, we obtain lim n ( n + a n ) log ( 1 + 1 n ) = 1 \lim_{n\to\infty} (n+a_n)\log\left(1+\frac{1}{n}\right)=1

Note that the power series for log ( 1 x ) -\log(1-x) is k = 1 x k k \sum_{k=1}^{\infty} \frac{x^k}{k} for x < 1 |x|<1 . Therefore:

lim n ( n + a n ) ( n 1 1 2 n 2 + 1 3 n 3 + O ( n 4 ) ) = 1 lim n ( n + a n ) ( n 1 1 2 n 2 ) = 1 lim n 1 2 n 1 + a n n 1 1 2 a n n 2 = 0 lim n a n n 1 = 1 2 n 1 a n = 1 2 \begin{aligned}\lim_{n\to\infty}(n+a_n)(n^{-1}-\frac12n^{-2}+\frac13n^{-3}+O(n^{-4}))&=1 \\ \lim_{n\to\infty}(n+a_n)(n^{-1}-\frac12n^{-2})&=1 \\ \lim_{n\to\infty}-\frac12n^{-1}+a_nn^{-1}-\frac12a_nn^{-2}&=0 \\ \lim_{n\to\infty}a_nn^{-1}&=\frac12n^{-1} \\ a_n&=\frac12\end{aligned}

EDIT: attempt at fixing massive error in the proof:

Due to the power series of log ( 1 x ) \log(1-x) , we see that as n n gets arbitrarily large, log ( 1 + n 1 ) \log(1+n^{-1}) gets arbitrarily close to n 1 1 2 n 2 n^{-1}-\frac12n^{-2} . Let a n a'_n be the solution to ( n + a n ) ( n 1 1 2 n 2 ) = 1 (n+a'_n)(n^{-1}-\frac12n^2)=1 . Then a n = 1 1 n 1 2 n 2 n = 1 ( 1 1 2 n 1 ) 1 n 1 2 n 2 = 1 2 1 1 1 2 n 1 a'_n=\frac1{\frac1n-\frac1{2n^2}}-n=\frac{1-(1-\frac12n^{-1})}{\frac1n-\frac1{2n^2}}=\frac12\frac{1}{1-\frac12n^{-1}}

Note that a n = 1 log ( 1 n 1 ) n a_n=\frac1{\log(1-n^{-1})}-n , so a n a n a_n\approx a'_n for arbitrarily large values of n n

Since lim n a n = lim n 1 2 1 1 1 2 n 1 \lim_{n\to\infty}a'_n=\lim_{n\to\infty}\frac12\frac{1}{1-\frac12n^{-1}} and lim n a n = a n \lim_{n\to\infty}a_n=a'_n , lim n a n = 1 2 \lim_{n\to\infty}a_n=\frac12

The second last equation isn't exactly correct. The LHS is a number (if it exists) and the RHS is a variable.

From your working, it need not be true that lim a n = 1 2 \lim a_n = \frac{1}{2} . For example, we could have a n = 1 3 a_n = \frac{1}{3} (or any other constant), and it will still satisfy the third last equation, namely

lim 1 2 n 1 + a n n 1 1 2 a n n 2 = 0. \lim - \frac{1}{2} n^{-1} + a_n n^{-1} - \frac{1}{2} a_n n^{-2} = 0.

You cannot simply 'cancel' when taking limits, especially since the limit is 0 and you are dividing by larger and larger numbers.

Chung Gene Keun - 7 years, 3 months ago

Log in to reply

Oh my goodness I can't believe I did that. I was really eager to reuse the proof that ( 1 + n 1 ) n + k (1+n^{-1})^{n+k} approaches e e fastest when k = 1 2 k=\frac12

Tim Ratigan - 7 years, 3 months ago

This proof is mathematically nt correct. you can't use limit and big O at the same equation on the same variable. what Chung Gene Keun said is another fallacy.

Nilabja Ray - 7 years, 3 months ago

Log in to reply

Thanks for the feedback, but actually I'm reasonably certain big O can be used in limits, unless you can prove me otherwise (please do, I prefer to know when I'm wrong).

However, I did realize that this was stupid reasoning, you can see my reply to Chung Keun.

Tim Ratigan - 7 years, 3 months ago

Log in to reply

Thanks for the feedback, but actually I'm reasonably certain big O can be used in limits, unless you can prove me otherwise (please do, I prefer to know when I'm wrong).

They can, in a way. This is really a moot question anyhow: all six uses of " lim n \lim_{n \rightarrow \infty} " need to be removed. Other than that, the main change to make is: don't suddenly drop the Big-O notation, it's the key to your solution.

Peter Byers - 7 years, 3 months ago
Alamin Sheikh
Feb 23, 2014

there is only one value between 0 & 1, that is 1/2 So, the solution is 1/2

The sequence 0 , 1 , 0 , 1 , 0 , 1 , 0, 1, 0, 1, 0, 1, \ldots does not have a limit.
The sequence 1 2 , 1 3 , 1 4 , \frac{1}{2} , \frac{1}{3} , \frac{1}{4}, \ldots tends to 0.

Why can't either of these work?

Chung Gene Keun - 7 years, 3 months ago

consider a(n) = (-1)^n.(1/3) ... this will also work but dosent converge... consider a(n) = (-1)^n (1/3)^n ... this will also work but converges to 0 so i think answer should be " does not exist'.

Robin Bhandari - 7 years, 3 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...