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 limit:

$$\lim_{n\to\infty} \left(1+\frac{1}{n}\right)\left(1+\frac{2}{n}\right)^{\frac{1}{2}}\cdots\left(1+\frac{n}{n}\right)^{\frac{1}{n}}$$

share|improve this question
    
I realise the question has been answered, but can someone please explain to me why one can't simply take the logarithm of the product, show that each term in that sum goes to zero, and conclude that the limit is 1? –  Daniel Littlewood Aug 17 '12 at 10:03
4  
@DanLitt A posteriori: because the limit is not 1. A priori: because no theorem says that the limit of some sums is the sum of the limits when the number of terms in the sums is not bounded. –  Did Aug 17 '12 at 10:14
    
Thank you, I needed the a priori statement –  Daniel Littlewood Aug 17 '12 at 10:19
4  
$1= \lim_{n \rightarrow \infty} 1 = \lim_{n \rightarrow \infty} (\sum_{i= 1}^{n} \frac{1}{n} )= \lim_{n \rightarrow \infty} \frac{1}{n} +\lim_{n \rightarrow \infty} \frac{1}{n} \ldots \lim_{n \rightarrow \infty} \frac{1}{n} =0+0+\ldots+0 = 0$ –  clark Aug 17 '12 at 11:00

2 Answers 2

up vote 19 down vote accepted

Note at the onset that $1+\frac{k}n\leqslant\mathrm e^{k/n}$ for every $k$ hence the $n$th product $P_n$ is such that $P_n\leqslant\mathrm e$, in particular, the sequence $(P_n)_{n\geqslant1}$ is bounded.

To show that $(P_n)_{n\geqslant1}$ actually converges and to identify its limit, note that, for every $n$, $$ \log(P_n)=\frac1n\sum\limits_{k=1}^nf\left(\frac{k}n\right), \qquad\text{with}\quad f(x)=\frac{\log(1+x)}x. $$ The function $f$ is continuous on $[0,1]$ (define $f(0)=1$) hence its Riemann sums converge to its integral and $P_n\to\mathrm e^\ell$ with $$ \ell=\int_0^1f(x)\mathrm dx=\int_0^1\left(\sum_{n\geqslant1}(-1)^{n+1}\frac{x^{n-1}}n\right)\mathrm dx=\sum_{n\geqslant1}\frac{(-1)^{n+1}}{n^2}=\frac{\pi^2}{12}. $$

share|improve this answer
    
Thanks for your nice answer. The integral way seems to be very useful here. –  Chris's sis Aug 17 '12 at 9:46
1  
in fact, on the second line you already have a Riemann sum and that is $-Li_{2}(-1)=\frac{\pi^2}{12}$. Oh, I missed this nice point :( –  Chris's sis Aug 17 '12 at 9:52
1  
+1: nice! (and nontrivial) –  Fabian Aug 17 '12 at 12:21

The limit is $+\infty$: Let's call $P_n = \displaystyle\prod_{i=1}^n(1+\frac{i}{n})^{\frac{1}{i}}$, and $S_n=\ln P_n = \displaystyle\sum_{i=1}^n\frac{1}{i}\ln(1+\frac{i}{n})$.

Then you have $S_n-S_{[\frac{n}{2}]}=\displaystyle\sum_{i\geq\frac{n}{2}}^n\frac{1}{i}ln(1+\frac{i}{n})\geq[\frac{n}{2}]\times\frac{1}{n}\times\ln(1+\frac{\frac{n}{2}}{n})\rightarrow \frac{1}{2}\ln(\frac{3}{2})\neq 0$ which means that $S_n$ diverges (if $(S_n)$ had a limit , $S_n - S_{[\frac{n}{2}]}$ would go to $0$).

share|improve this answer
    
it's not $\infty$. See did's answer. –  Chris's sis Aug 17 '12 at 10:04
5  
a short practical computational checkout (for instance Pari/GP) should have made that wrong answer not appear... –  Gottfried Helms Aug 17 '12 at 10:10

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.