I tried using Stirling's approximation and d'Alambert's ratio test but can't get the limit. Could someone show how to evaluate this limit?
-
Well, this follows from Stirling. So you should maybe show us what you tried and say why it didn't work... – David C. Ullrich Aug 09 '15 at 21:04
-
Try Stirling again. – Thomas Andrews Aug 09 '15 at 21:13
6 Answers
Without Stirling:
Let $u_n=\frac{\sqrt[n]{n!}}{n}$. We have that $$\ln(u_n)=\frac{1}{n}\sum_{k=1}^n\ln\left(\frac{k}{n}\right)$$ and thus $$\lim_{n\to\infty }\ln(u_n)=\int_0^1 \ln(x)\mathrm{d}x=-1.$$ Since $x\mapsto e^x$ is continuous at $x=-1$, $$\lim_{n\to\infty }u_n=e^{-1}=\frac{1}{e}.$$
- 55,662
-
Yeah, you have to deal with the fact that $\int_{0}^1 \ln(x),dx$ is an improper integral, but this is a neat answer. – Thomas Andrews Aug 09 '15 at 21:23
-
@Surb: $\ln(x)$ at $0$ is negative infinity, so formally you should have a limit of $m$ which tends to $0^+$ in the integrand. – Eemil Wallin Aug 09 '15 at 21:30
-
@surb The fix is trivial ... just interpret the sum as the Riemann sum that evaluates the integrand at the right side of each subinterval. Then the lower limit is $1/n$ – Mark Viola Aug 09 '15 at 21:36
-
-
@OussamaBoussif: No it's $\ln(x)$. By definition, $$\lim_{n\to\infty }\frac{1}{n}\sum_{k=1}^n f\left(\frac{k}{n}\right)=\int_0^1 f(x)dx.$$ – Surb Aug 09 '15 at 22:16
-
yes I know but maybe you got a different answer when you changed the limits of product – Oussama Boussif Aug 09 '15 at 22:18
This is an alternate solution for those who don't know stirling approximation yet like me
$$ \lim _{ n\longrightarrow +\infty }{ \frac { \sqrt [ n ]{ n! } }{ n } } \quad =\quad \lim _{ n\longrightarrow +\infty }{ \frac { \sqrt [ n ]{ \prod _{ k=0 }^{ n-1 }{ (n-k) } } }{ n } } \\ \qquad \qquad \qquad \qquad =\quad \lim _{ n\longrightarrow +\infty }{ \frac { \sqrt [ n ]{ \prod _{ k=0 }^{ n-1 }{ n(1-\frac { k }{ n } ) } } }{ n } } \\ \qquad \qquad \qquad \qquad =\quad \lim _{ n\longrightarrow +\infty }{ \frac { \sqrt [ n ]{ { n }^{ n }\prod _{ k=0 }^{ n-1 }{ (1-\frac { k }{ n } ) } } }{ n } } \\ \qquad \qquad \qquad \qquad =\quad \lim _{ n\longrightarrow +\infty }{ \frac { n\sqrt [ n ]{ \prod _{ k=0 }^{ n-1 }{ (1-\frac { k }{ n } ) } } }{ n } } \\ \qquad \qquad \qquad \qquad =\quad \lim _{ n\longrightarrow +\infty }{ { \left( \prod _{ k=0 }^{ n-1 }{ \left( 1-\frac { k }{ n } \right) } \right) }^{ \frac { 1 }{ n } } } \\ \qquad \qquad \qquad \qquad =\quad \lim _{ n\longrightarrow +\infty }{ { e }^{ \frac { 1 }{ n } \sum _{ k=0 }^{ n-1 }{ \ln { \left( 1-\frac { k }{ n } \right) } } } } $$
Now we will evaluate that infinite sum:
$$ \underset { n\longrightarrow +\infty }{ lim } \frac { 1 }{ n } \sum _{ k=0 }^{ n-1 }{ \ln { \left( 1-\frac { k }{ n } \right) } } \quad =\underset { n\longrightarrow +\infty }{ lim } \quad \frac { 1-0 }{ n } \sum _{ k=0 }^{ n-1 }{ \ln { \left( 1-\frac { k }{ n } \right) } } \\ \qquad \qquad \qquad \qquad \qquad =\quad \int _{ 0 }^{ 1 }{ ln(1-x)\quad dx } \\ \qquad \qquad \qquad \qquad \qquad =\quad { \left[ x\ln { (1-x) } \right] }_{ 0 }^{ 1 }+\int _{ 0 }^{ 1 }{ \frac { x }{ 1-x } dx } \\ \qquad \qquad \qquad \qquad \qquad =\quad { \left[ x\ln { (1-x) } \right] }_{ 0 }^{ 1 }+{ \left[ -\ln { (1-x)-x } \right] }_{ 0 }^{ 1 }\\ \qquad \qquad \qquad \qquad \qquad =\quad { \left[ (x-1)\ln { (1-x)-x } \right] }_{ 0 }^{ 1 }\quad =\quad -1 $$
So your limit is: ${e}^{-1}$
- 3,085
Any version of Stirling's formula, including approximate ones that only give the order of magnitude without the right constant or power of $n$, will show that limit is $1/e$. The dominant piece of the formula is $(n/e)^n$ and any corrections to that die out when taking $n$-th roots.
- 2,434
Use equivalents: $$\frac{\sqrt[n]{n!}}n\sim_{\infty}\frac{\bigl(\sqrt{2\pi n}\bigr)^{\tfrac 1n}}{n}\cdot\frac n{\mathrm{e}}=\frac 1{\mathrm{e}}\bigl({2\pi n}\bigr)^{\tfrac 1{2n}}$$ Now $\;\ln\bigl({2\pi n}\bigr)^{\tfrac 1{2n}}=\dfrac{\ln\pi+\ln 2n}{2n}\xrightarrow[n\to\infty]{}0$, hence $$\frac{\sqrt[n]{n!}}n\sim_{\infty}\frac 1{\mathrm{e}}. $$
- 175,478
-
Of course there are a few devilish little details hidden in that "use equivalents". In particular, the fact that if $A_n\sim B_n$ then $A_n^{1/n}\sim B_n^{1/n}$. Which of course is not hard, but it needs to be done; for example $A_n\sim B_n$ does not imply $A_n^n\sim B_n^n$... – David C. Ullrich Aug 09 '15 at 21:19
-
Yes I skipped some details. Some job must be left to the O.P., don't you think? – Bernard Aug 09 '15 at 21:22
Stirling's Formula is
$$n!=\sqrt{2\pi\,n}\left(\frac{n}{e}\right)^n\left(1+O\left(\frac1n\right)\right)$$
From this it is easy to see that
$$(n!)^{1/n}=(2\pi\,n)^{1/2n}\frac{n}{e}\left(1+O\left(\frac{1}{n^2}\right)\right)$$
whereupon dividing by $n$ and letting $n\to \infty$ yields
$$\bbox[5px,border:2px solid #C0A000]{\lim_{n\to \infty}\frac{(n!)^{1/n}}{n}=e^{-1}}$$
- 179,405
$$\frac { \sqrt [ n ]{ n! } }{ n } =\sqrt [ n ]{ \frac { n! }{ { n }^{ n } } } =\sqrt [ n ]{ { x }_{ n } } \\ x_{ n }=\frac { n! }{ { n }^{ n } } \\$$ Hence $\forall n\epsilon \quad N,{ x }_{ n }>0$ ,this formula is true
$$\lim _{ n\rightarrow \infty }{ \sqrt [ n ]{ { x }_{ n } } =\lim _{ n\rightarrow \infty }{ \frac { { x }_{ n } }{ { x }_{ n-1 } } } } $$
$$ \lim _{ n\rightarrow \infty }{ \frac { { x }_{ n-1 } }{ { x }_{ n } } = } \lim _{ n\rightarrow \infty }{ { \left( 1+\frac { 1 }{ n-1 } \right) }^{ n-1 } } =\frac { 1 }{ e } $$
- 21,578
-
-
-
-
why to be so angry? I can prove it,it might not be formula as i said,my english so poor,sorry for it $\lim { n\rightarrow \infty }{ \sqrt [ n ]{ { x }{ n } } =\lim { n\rightarrow \infty }{ \sqrt [ n ]{ { x }{ 1 }\cdot \frac { { x }{ 2 } }{ { x }{ 1 } } \cdot \frac { { x }{ 3 } }{ { x }{ 2 } } \cdot ...\cdot \frac { { x }{ n } }{ { x }{ n-1 } } } =\lim { n\rightarrow \infty }{ \frac { { x }{ n } }{ { x }_{ n-1 } } } } } $, – haqnatural Aug 09 '15 at 21:46
-
Actually, I was not being angry at all, I was only asking that you provide the reasoning that shows that the formula is correct. Please no worry. But this development here seems only to prove that $x_n=x_n$. – Mark Viola Aug 09 '15 at 22:18