Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Compute the following limit:

$$\lim_{n\to\infty}\frac{{\left(1^1 \cdot 2^2 \cdot3^3\cdots n^n\right)}^\frac{1}{n^2}}{\sqrt{n}} $$

I'm interested in almost any approaching way for this limit. Thanks.

share|improve this question
2  
do you know the answer? take the limit equal to some k and then log on both the sides. –  sukunrt Jun 23 '12 at 20:28
2  
+1. I find that all your questions, in general, to be interesting and intriguing. –  user17762 Jun 23 '12 at 20:34
    
I wonder who is that troll, diligently downvoting your questions... –  userNaN Jun 23 '12 at 20:40

2 Answers 2

up vote 21 down vote accepted

Let's begin $$ \lim\limits_{n\to\infty}\frac{\left(\prod\limits_{k=1}^n k^k\right)^{\frac{1}{n^2}}}{\sqrt{n}}= \lim\limits_{n\to\infty}\exp\left(\frac{1}{n^2}\sum\limits_{k=1}^n k\log k - \frac{1}{2}\log n\right)= $$ $$ \lim\limits_{n\to\infty}\exp\left(\frac{1}{n^2}\sum\limits_{k=1}^n k\log\left(\frac{k}{n}\right)+\frac{1}{n^2}\sum\limits_{k=1}^n k\log n - \frac{1}{2}\log n\right)= $$ $$ \lim\limits_{n\to\infty}\exp\left(\sum\limits_{k=1}^n \frac{k}{n}\log\left(\frac{k}{n}\right)\frac{1}{n}+\frac{1}{2}\log n\left(\frac{n^2+n}{n^2}-1\right)\right)= $$ $$ \exp\left(\lim\limits_{n\to\infty}\sum\limits_{k=1}^n \frac{k}{n}\log\left(\frac{k}{n}\right)\frac{1}{n}+\frac{1}{2}\lim\limits_{n\to\infty}\frac{\log n}{n}\right)= $$ $$ \exp\left(\int\limits_{0}^1 x\log x dx\right)=\exp\left(-1/4\right) $$ And now we are done!

share|improve this answer
1  
very fast and elegant at the same time! Thanks! –  Chris's sis Jun 23 '12 at 20:33
    
You are welcome! –  userNaN Jun 23 '12 at 20:35

$$ \frac1{n^2}\sum_{k=1}^nk\log(k)-\frac12\log(n)=\frac1{n}\sum_{k=1}^n\frac{k}n\log\left(\frac{k}n\right)+\frac12\frac{\log(n)}n=\int_0^1x\log(x)\mathrm dx+o(1) $$


Edit: Per request, a solution without integrals, using only the elementary version of Stirling's approximation. Let $s_n=\displaystyle\sum_{k=1}^nk\log(k)$. Then, $$ \sum_{k=1}^n\log(k!)=\sum_{k=1}^n\sum_{i=1}^k\log(i)=\sum_{i=1}^n(n-i+1)\log(i)=(n+1)\log(n!)-s_n. $$ On the other hand, Stirling's approximation in its simplest form is $\log(k!)=k\log(k)-k+r_k$ with $r_k=O(\log k)$, which yields $$ \sum_{k=1}^n\log(k!)=\sum_{k=1}^nk\log(k)-\sum_{k=1}^nk+\sum_{k=1}^nr_k=s_n-\frac12n(n+1)+t_n,\quad t_n=\sum_{k=1}^nr_k. $$ Comparing these two expressions and using once more $\log(n!)=n\log(n)-n+r_n$, one gets $$ 2s_n=n(n+1)\log n-n(n+1)+(n+1)r_n+\frac12n(n+1)-t_n, $$ hence $2s_n-n^2\log n=-\frac12n^2+u_n$ with $$ u_n=n\log(n)-\frac12n+(n+1)r_n-t_n. $$ Since $r_n=O(\log n)$, $t_n=O(n\log n)$ and each term in $u_n$ is $O(n\log n)$, hence $$ \frac{s_n}{n^2}-\frac{\log n}2=-\frac14+\frac{u_n}{2n^2}=-\frac14+O\left(\frac{\log n}{n}\right)\to-\frac14. $$

share|improve this answer
    
thanks for your solution. –  Chris's sis Jun 23 '12 at 20:34

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.