3
$\begingroup$

There is the following limit, I would like to calculate:

$\lim_{n\rightarrow\infty}\frac{n!}{\left(n+1/6\right)!}$

I tried to use the Stirling approaximation formula

$n!\approx\sqrt{2\pi n}\left(\frac{n}{e}\right)^{n}$

After the substituion I have got a relatively complex formula. I suppose that it may be solved by the Hospital's rule...

Is it the right method for the limit computation, if we don't want to use the Gamma function?

Thanks for your help...

$\endgroup$
3
  • 4
    $\begingroup$ The problem is how do you define $(n+\frac{1}{6})!$ without the Gamma function? $\endgroup$
    – Lord Soth
    Commented Jul 27, 2013 at 18:49
  • 1
    $\begingroup$ Well, if one uses Stirling's approximation, one needs not make any explicit mention of the gamma function. $\endgroup$ Commented Jul 27, 2013 at 18:51
  • 4
    $\begingroup$ @BabyDragon Even if you do not say "Gamma," once you write $(n+1/6)!$, everybody will know that you are using the Gamma function, you have nowhere to hide. $\endgroup$
    – Lord Soth
    Commented Jul 27, 2013 at 18:57

2 Answers 2

4
$\begingroup$

If you use Stirling's approximation you get $$\frac{n!}{\left(n+1/6\right)!}\approx\frac{\sqrt{2\pi n}\left(\frac{n}{e}\right)^{n}}{\approx\sqrt{2\pi (n+1/6)}\left(\frac{n+1/6}{e}\right)^{n+1/6}}=\frac{\sqrt{2\pi n}}{\sqrt{2\pi (n+1/6)}}\frac{1}{(\frac{n+1/6}{e})^{1/6}}(\frac{n}{n+1/6})^n$$. Now note that $$\frac{\sqrt{2\pi n}}{\sqrt{2\pi (n+1/6)}}\approx \frac{\sqrt{n}}{\sqrt{n+1/6}}\approx 1$$ and $$\frac{1}{(\frac{n+1/6}{e})^{1/6}}\approx 0$$ Now let us deal with the last bit of the product. For this we take the logarithm to get $$n(\ln(n)-\ln(n+1/6))=\frac{\ln(n)-\ln(n+1/6)}{1/n}$$ Hitting this with L'Hopital's, we get $$\lim_{n\to\infty}\frac{\ln(n)-\ln(n+1/6)}{1/n}=\lim_{n\to\infty}\frac{\frac{1}{n}-\frac{1}{n+1/6}}{-1/n^2}=\lim_{n\to\infty}-\frac{n^2/6}{n(n+1/6)}=-\frac{1}{6}.$$ Therefore, $\lim_{n\to\infty}(\frac{n}{n+1/6})^n=e^{-1/6}$. Putting all of this together, we get $$\lim_{n\to\infty}\frac{n!}{\left(n+1/6\right)!}=0$$

$\endgroup$
4
  • 2
    $\begingroup$ You lost a factor of $\frac16$, the limit of $\left(\frac{n}{n+1/6}\right)^n$ is $e^{-1/6}$. Of course, the final limit of $0$ is unchanged. $\endgroup$ Commented Jul 27, 2013 at 20:28
  • $\begingroup$ @DanielFischer Thank you. It is fixed. $\endgroup$ Commented Jul 27, 2013 at 20:29
  • $\begingroup$ Wow, that was fast. $\endgroup$ Commented Jul 27, 2013 at 20:30
  • 1
    $\begingroup$ What can I say, I lead a sad life. $\endgroup$ Commented Jul 27, 2013 at 20:31
1
$\begingroup$

The derivation of Stirling's formula not relying on the gamma function depends on the fact that $n\in\mathbb{N}$. By using Stirling's formula for $n\notin\mathbb{N}$ we are analytically extending the formula to numbers for which the original derivation fails. As mentioned in the comments, secretly we are using the gamma function.

In any case, using Stirling's formula we find $$\begin{eqnarray*} \log\frac{n!}{(n+a)!} &\sim& \log \sqrt{2\pi n}(n/e)^n - \log \sqrt{2\pi(n+a)}((n+a)/e)^{n+a} \\ &=& \left(\log\sqrt{2\pi}+\frac{1}{2}\log n + n\log n - n\right) \\ && - \left(\log\sqrt{2\pi}+\frac{1}{2}\log(n+a) + (n+a)\log(n+a) - (n+a)\right) \\ &=& -\frac{1}{2}\log(1+a/n) - n\log(1+a/n) - a\log n - a\log(1+a/n) + a\\ &\sim& -a\log n, \end{eqnarray*}$$ where in the last step we use $\log(1+x)=x+O(x^2)$ for $x\ll 1$ and neglect terms of order $1/n$. Therefore, $n!/(n+a)! \sim 1/n^a\,(n\to\infty)$ and so, if $a>0$, $$\lim_{n\to\infty}\frac{n!}{(n+a)!} = 0.$$

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .