I've attempted to solve this $\lim_{n\to\infty}$ $\frac 1{(n!)^\frac 1{n}}$ via converging 1/n to zero, but somehow it's incorrect. Can someone explain what i'm doing wrong?
-
2use stirling approximation to write n! – Anindya Prithvi Oct 08 '20 at 04:24
-
Thank's for your reply, but i know that it can be solved without stirling approximation. I mean, i can't rely on it every time – Ramil Ahmedov Oct 08 '20 at 04:27
-
use logarithms and convert it into an integral – Anindya Prithvi Oct 08 '20 at 04:30
-
Also tried and it worked, but still there is solution via squeeze theorem. – Ramil Ahmedov Oct 08 '20 at 04:32
-
1Then you have elaborate it in your question well. you must be getting the answer as 1(using squeeze thm), right? – Anindya Prithvi Oct 08 '20 at 04:33
-
You' re right, next time i'll write it down in description. The correct answer is 0 – Ramil Ahmedov Oct 08 '20 at 04:35
-
Use $$\frac{x^{x}}{e^{x}}<x!<x^{x}$$ for $x>1$ – Mourad Oct 08 '20 at 04:44
4 Answers
Hint: $$0<\frac 1{(n!)^\frac 1{n}}=\sqrt[n]{\frac{1}{n!}}<\frac{1+\frac{1}{2}+\cdots+\frac{1}{n}}{n}\to 0.$$ In general, we have: if $a_n\to a$, then $$\lim_{n\to\infty}\frac{a_1+a_2+\cdots+a_n}{n}=a.$$

- 7,203
-
Hi, thank you for your reply, however this approach is not obvious. How can we prove the part with < inequality? – Ramil Ahmedov Oct 09 '20 at 04:03
-
the inequality used is AM-GM inequality! And you can refer https://math.stackexchange.com/questions/691807/proofs-of-am-gm-inequality – Riemann Oct 09 '20 at 09:04
-
AM-GM inequality is an elementary inequality even for high school students! You should master it (learn to use it). – Riemann Oct 09 '20 at 09:08
The limit is $0$, and the following very hands-on argument suffices: for any positive integer $k \le n$ we have
$$n! \ge k^{n-k}$$
(we look at $1 \cdot 2 \cdot \dots \cdot n$ and just ignore the terms less than or equal to $k$ and replace all the terms bigger than $k$ with $k$), which gives
$$\sqrt[n]{n!} \ge k^{1 - \frac{k}{n}}.$$
Taking, for example, $k = \lfloor \frac{n}{2} \rfloor$ gives
$$\sqrt[n]{n!} \ge \left\lfloor \frac{n}{2} \right\rfloor^{\frac{1}{2}}$$
so $\lim_{n \to \infty} \sqrt[n]{n!} = \infty$ and $\lim_{n \to \infty} \frac{1}{\sqrt[n]{n!}} = 0$. Of course much tighter bounds are available, and even quite exact asymptotics via Stirling's formula, but it's worth knowing what can be proven in a very elementary way.

- 419,620
-
I like this one, but how can we proove that n! is actually more than k^n-k – Ramil Ahmedov Oct 09 '20 at 04:08
-
@Amil: I gave the argument in parentheses: you look at the product $1 \cdot 2 \cdot \dots \cdot n$, ignore the terms less than or equal to $k$, and bound the terms greater than or equal to $k$ by $k$. It's very simple. – Qiaochu Yuan Oct 09 '20 at 06:10
$$y=\lim_{n\to\infty}\frac 1{(n!)^\frac 1{n}}$$
$$\ln y=\lim_{n\to\infty}-\frac{1}{n}\ln(n!)$$
$$\ln y=\lim_{n\to\infty}-\frac{1}{n}\left({\ln1+\ln2+\cdots\ln n}\right)$$
$$\ln y=\lim_{n\to\infty}-\frac{1}{n}\left({[\ln\frac{1}{n}+\ln\frac{2}{n}+\cdots\ln\frac{n}{n}]+n\ln n}\right)$$
$$\ln y=\lim_{n\to\infty}\left({\int_0^1\ln x dx-\ln n}\right)$$
$$\ln y=\lim_{n\to\infty}\left({-1-\ln n}\right)\to-\infty$$
$$y=e^{-\infty}\to0$$

- 4,055
-
How did you convert the rienman sum into integral without a $ \frac{1}{n}..?$ – tryst with freedom Oct 20 '20 at 10:21
-
-
Use the common limit $$\lim_{n \to \infty} \left( \frac{n^n}{n!}\right)^\frac 1n = \lim_{n \to \infty} n \times \frac{1}{(n!)^\frac1n} = e$$ It's easy to see that $$\lim_{n \to \infty} \frac{1}{(n!)^\frac1n} = \lim_{n \to \infty} \frac{e}{n} = 0$$
And that's the final answer.
It's easier to do this via Stirling's approximation, as $n! = \sqrt{2\pi n}\left( \frac ne\right)^n$ and the nth root of $1/n!$ is equal to $e/n$, which tends to 0.

- 31,845

- 4,345
- 11
- 30