Find$$\lim_{n \rightarrow \infty}\left(\frac{(2n)!}{n!n^n} \right)^{1/n}$$ is there some trick in this questions. seems it must simplify to something but I am unable to solve it.
-
@ClementC. unfortunately not too fancy solution but for a student of grade 12 – user6775 Jun 09 '17 at 11:34
-
Then, if you know about Riemann sums (my previous comment asked about Stirling's approximation), @labbhattacharjee's answer is perfect. – Clement C. Jun 09 '17 at 11:36
-
The AM-GM inequality easily gets you that the sequence lies in the range $[1, 1.5]$. – Jack M Jun 09 '17 at 15:32
7 Answers
$$\begin{eqnarray*}\lim_{n\to +\infty}\left(\frac{(2n)!}{n!n^n}\right)^{\frac{1}{n}}&=& \lim_{n\to +\infty}\left[\prod_{k=1}^{n}\left(1+\frac{k}{n}\right)\right]^{\frac{1}{n}}\\&=&\exp\lim_{n\to +\infty}\frac{1}{n}\sum_{k=1}^{n}\log\left(1+\frac{k}{n}\right)\\&=&\exp\int_{0}^{1}\log(1+x)\,dx\\&=&\exp\left(\log(4)-1\right)=\color{red}{\frac{4}{e}}\end{eqnarray*}$$ by employing the continuity of the exponential/logarithm function and a Riemann sum.
- 353,855
Using Stirling's approximation, just for the sake of it and for reference (and to give an alternative approach to that of the excellent answer by lab bhattacharjee). Caveat: overly detailed.
$$ n! \operatorname*{\sim}_{n\to\infty} \sqrt{2\pi n}\left(\frac{n}{e}\right)^n \tag{Stirling's approximation} $$ yields $$ \frac{(2n)!}{n!n^n} \operatorname*{\sim}_{n\to\infty} \frac{\sqrt{4\pi n}\left(\frac{2n}{e}\right)^{2n}}{\sqrt{2\pi n}\left(\frac{n}{e}\right)^n n^n} = \frac{\sqrt{2}\left(\frac{2}{e}\right)^{2n}}{\left(\frac{1}{e}\right)^n}= \sqrt{2}\left(\frac{4}{e}\right)^{n} $$ i.e. $$ \frac{(2n)!}{n!n^n} = \sqrt{2}\left(\frac{4}{e}\right)^{n} + o\left(\left(\frac{4}{e}\right)^{n}\right). $$
From there, $$\begin{align} \left(\frac{(2n)!}{n!n^n}\right)^{1/n} &= \exp\left(\frac{1}{n}\ln\left(\frac{(2n)!}{n!n^n}\right)\right) = \exp\left(\frac{1}{n}\ln\left(\sqrt{2}\left(\frac{4}{e}\right)^{n} + o\left(\left(\frac{4}{e}\right)^{n}\right)\right)\right) \\ &= \exp\left(\frac{1}{n}\ln\left(\sqrt{2}\left(\frac{4}{e}\right)^{n}\right)+\frac{1}{n}\ln(1+o(1)\right) \\ &= \exp\left(\frac{1}{n}\ln\left(\left(\frac{4}{e}\right)^{n}\right)+\frac{1}{n}\ln(\sqrt{2})+\frac{1}{n}\ln(1+o(1)\right) \\ &= \exp\left(\ln\left(\frac{4}{e}\right)+o(1)+o(1)\right) \\ &= \left(\frac{4}{e}\right)e^{o(1)} \\ &\xrightarrow[n\to\infty]{} \boxed{\frac{4}{e}}. \end{align}$$
- 67,323
-
Note: the above works even if you don't know the exact constant $\sqrt{2\pi}$ in Stirling's approximation, i.e. only know the (much simpler to establish) asymptotics $$n!\sim_{n\to\infty} C\sqrt{n}\left(\frac{n}{e}\right)^n$$ for some unspecified $C>0$. (Since anyway, that constant gets cancelled immediately in the argument.) – Clement C. Jun 09 '17 at 15:25
Just use the following result:
If $a_{n} $ is a sequence of positive terms and $a_{n+1}/a_{n}\to L$ then $a_{n} ^{1/n}\to L$.
The above is easily proved by taking logs and applying Cesaro-Stolz. For the current question we take $a_{n} = (2n)!/(n!n^{n})$ and then we can see that $$\frac{a_{n+1}}{a_{n}}=\frac{(2n+1)(2n+2)}{(n+1)^{2}}\cdot\dfrac{1}{\left(1+\dfrac{1}{n}\right)^{n}}\to\frac{4}{e}$$ so the desired limit is $4/e$.
On request of user "S.H.W" (via comment) I provide proof of the result used in this answer. Let $b_{n} =\log a_{n} $ then $b_{n} $ is well defined as $a_{n} $ is positive. Now we can apply Cesaro-Stolz on $b_{n} /n$ to get the relation $$\lim_{n\to\infty} \frac{b_{n}} {n} =\lim_{n\to\infty} \frac{b_{n+1}-b_{n}}{n+1-n}$$ provided the limit on right hand side exists finitely or infinitely. This means that $$\lim_{n\to\infty} \log(a_{n} ^{1/n})=\lim_{n\to\infty}\log(a_{n+1}/a_{n})$$ and on exponentiating (which is a continuous operation) we get $$\lim_{n\to\infty} a_{n} ^{1/n}=\lim_{n\to\infty}\frac{a_{n+1}}{a_{n}}$$ provided the limit on right hand exists finitely or infinitely. And thus the proof is completed via Cesaro-Stolz.
Most textbooks provide proofs without the use of Cesaro-Stolz and logarithms and you should have a look at that proof also because it is typical of the proofs used in the theory of limits of sequences. Roughly the idea is that if the ratio $a_{n+1}/a_{n}$ lies between $L-\epsilon $ and $L+\epsilon$ for all $n\geq m$ then $a_{n} /a_{m} $ lies between $(L-\epsilon) ^{n-m} $ and $(L+\epsilon) ^{n-m} $.
- 87,309
-
Your answer is brilliant , Can you explain how did you conclude that statement from Cesaro-Stolz ? I tried but didn't get result . – S.H.W Jun 10 '17 at 19:52
-
HINT:
$$\ln f=\lim_{n\to\infty}\dfrac1n\ln\dfrac{(2n)!}{n^n\cdot n!}$$
$$\dfrac{(2n)!}{n^n\cdot n!}=\dfrac{\prod_{r=1}^n(r+n)}{n^n}=\prod_{r=1}^n\left(1+\dfrac rn\right)$$
Now use the idea of How to represent this limit?
- 274,582
Note that $$ \left(\frac{(2n)!}{n!n^n} \right)^{1/n}=\left(\binom{2n}{n}\frac{n!}{n^n}\right)^{1/n}. $$ Now, thanks to Wallis product, we have $$ \binom{2n}{n}\sim \frac{4^n}{\sqrt{\pi n}}. $$ On a similar line, considering Stirling approximation of $n!$, we obtain $$ \left(\frac{(2n)!}{n!n^n} \right)^{1/n}\sim \left(\frac{4^n}{\sqrt{\pi n}}\cdot \frac{\sqrt{2\pi n}(n/e)^n}{n^n}\right)^{1/n}=\frac{4}{e}(\sqrt{2})^{1/n}\to \frac{4}{e}. $$
- 15,423
- 3
- 24
- 57
-
Err, are you sure? The limit is $\exp \int_0^1 \ln(1+x) , dx = \frac{4}{e}$. Your inequality gives an upper bound of $$\left(\frac{(2n)!}{n!n^n} \right)^{1/n}< \frac{2}{\sqrt{n}^{1/n}}$$ which does not converge to $0$. – Clement C. Jun 09 '17 at 11:41
-
1I just made the limit of the exponent $\frac{1}{n}\log \frac{1}{\sqrt{n}}$ and forgot about the base $e$. Ops, I delete it :) Thanks Clement – Paolo Leonetti Jun 09 '17 at 11:45
-
Small remark: After employing Sitrling's approximation, it is not entirely clear (for someone who didn't see it before, as seems to be the case for the OP) that the step "$a_n \sim b_n$ yields $a_n^{1/n}\sim b_n^{1/n}$" is legitimate (it is here). – Clement C. Jun 09 '17 at 12:02
-
1
$\newcommand{\bbx}[1]{\,\bbox[15px,border:1px groove navy]{\displaystyle{#1}}\,} \newcommand{\braces}[1]{\left\lbrace\,{#1}\,\right\rbrace} \newcommand{\bracks}[1]{\left\lbrack\,{#1}\,\right\rbrack} \newcommand{\dd}{\mathrm{d}} \newcommand{\ds}[1]{\displaystyle{#1}} \newcommand{\expo}[1]{\,\mathrm{e}^{#1}\,} \newcommand{\ic}{\mathrm{i}} \newcommand{\mc}[1]{\mathcal{#1}} \newcommand{\mrm}[1]{\mathrm{#1}} \newcommand{\pars}[1]{\left(\,{#1}\,\right)} \newcommand{\partiald}[3][]{\frac{\partial^{#1} #2}{\partial #3^{#1}}} \newcommand{\root}[2][]{\,\sqrt[#1]{\,{#2}\,}\,} \newcommand{\totald}[3][]{\frac{\mathrm{d}^{#1} #2}{\mathrm{d} #3^{#1}}} \newcommand{\verts}[1]{\left\vert\,{#1}\,\right\vert}$
\begin{align} &\lim_{n \to\ \infty}\ln\pars{\braces{\bracks{2n}! \over n!\,n^{n}}^{1/n}} = \lim_{n \to\ \infty} {\ln\pars{\bracks{2n}!} - \ln\pars{n!} - n\ln\pars{n} \over n} \\[5mm] = &\ \lim_{n \to\ \infty} \left\{\vphantom{\Large A}\bracks{\vphantom{\large A}\ln\pars{2} + \ln\pars{n + 1} + \ln\pars{2n + 1}} - \ln\pars{n + 1}\right. \\[2mm] &\ \left.\phantom{\,\lim_{n \to \infty}} - \bracks{\vphantom{\large A}n\ln\pars{n + 1} + \ln\pars{n + 1} - n\ln\pars{n}}\right\} \\[5mm] = &\ \lim_{n \to \infty}\left[2\ln\pars{2} + \ln\pars{n} + \ln\pars{1 + {1 \over 2n}} - n\ln\pars{n} - n\ln\pars{1 + {1 \over n}} - \ln\pars{n} - \ln\pars{1 + {1 \over n}}\right. \\[2mm] &\ \left.\phantom{\lim_{n \to \infty}}+ n\ln\pars{n}\right] \\[5mm] = &\ \ln\pars{4}\ - \ \underbrace{\lim_{n \to \infty}\bracks{n\ln\pars{1 + {1 \over n}}}} _{\ds{\to\ 1\ \mbox{as}\ n\ \to\ \infty}}\ +\ \underbrace{\lim_{n \to \infty} \bracks{\ln\pars{1 + {1 \over 2n}} - \ln\pars{1 + {1 \over n}}}} _{\ds{\to\ 0\ \mbox{as}\ n\ \to\ \infty}}\ =\ \bbx{\ln\pars{4} - 1} \end{align}
$$ \lim_{n \to\ \infty}\bracks{\pars{2n}! \over n!\,n^{n}}^{1/n} = \exp\pars{\ln\pars{4} - 1} = \bbox[#ffe,10px,border:1px dotted navy]{\ds{4 \over \expo{}}} $$
- 89,464
The easiest way to solve the limit with high school knowledge is by realizing that the limit of the n-th root of a succession equals the limit of the ratio of two consecutive terms. EDIT: provided the limit of the ratio exists.
- 53
-
1Is that true without further assumptions, though? Consider the sequence $(a_n)_n$defined by $$a_n \stackrel{\rm def}{=} \frac{2+(-1)^n}{2^n}$$ – Clement C. Jun 09 '17 at 15:10
-
-
This is exactly what my answer does and with all details. Perhaps you had not seen my answer and posted yours. Also I am not sure if the result you are talking of belongs to high school syllabus (students upto age 17). – Paramanand Singh Jun 11 '17 at 08:08