14

Calculate the sum of series:

$$\sum_{n=1}^\infty \ln\left(\frac{n(n+2)}{(n+1)^2}\right)$$

I tried to spread this logarithm, but I'm not seeing any method for this exercise.

Yas
  • 283

6 Answers6

30

Note $$\ln\left(\frac{n(n+2)}{(n+1)^2}\right)=\ln\left(\frac{\frac{n}{n+1}}{\frac{n+1} {n+2}}\right)=\ln\left(\frac{n}{n+1}\right)-\ln\left(\frac{n+1}{n+2}\right)$$ Thus $$\sum_{n=1}^\infty \ln\left(\frac{n(n+2)}{(n+1)^2}\right)=\sum_{n=1}^\infty \left[\ln\left(\frac{n}{n+1}\right)-\ln\left(\frac{n+1}{n+2}\right)\right]$$ This is a telescoping series. Therefore $$\sum_{n=1}^\infty \ln\left(\frac{n(n+2)}{(n+1)^2}\right)=-\ln(2)$$

  • 1
    Maybe worth pointing out that $\ln\left(\frac{n(n+2)}{(n+1)^2}\right)=\ln\left(\frac{\frac{n}{n+1}}{\frac{n+1} {n+2}}\right)$ only for $n \neq -1, -2$, but since the summation starts at $1$ that's ok. – Chris Bouchard Jan 05 '17 at 14:59
23

An overkill. Since holds $$\prod_{n\geq0}\frac{\left(n+a\right)\left(n+b\right)}{\left(n+c\right)\left(n+d\right)}=\frac{\Gamma\left(c\right)\Gamma\left(d\right)}{\Gamma\left(a\right)\Gamma\left(b\right)},\, a+b=c+d $$ and this can be proved using the Euler's definition of the Gamma function, we have $$\sum_{n\geq1}\log\left(\frac{n\left(n+2\right)}{\left(n+1\right)^{2}}\right)=\log\left(\prod_{n\geq0}\frac{\left(n+1\right)\left(n+3\right)}{\left(n+2\right)\left(n+2\right)}\right)=\log\left(\frac{\Gamma\left(2\right)\Gamma\left(2\right)}{\Gamma\left(1\right)\Gamma\left(3\right)}\right)=\color{red}{\log\left(\frac{1}{2}\right)}.$$

Marco Cantarini
  • 33,062
  • 2
  • 47
  • 93
8

In another, more straight, way: $$ \begin{gathered} \sum\limits_{1\, \leqslant \,n} {\ln \left( {\frac{{n\left( {n + 2} \right)}} {{\left( {n + 1} \right)^{\,2} }}} \right)} = \ln \left( {\prod\limits_{1\, \leqslant \,n} {\frac{{n\left( {n + 2} \right)}} {{\left( {n + 1} \right)^{\,2} }}} } \right) = \hfill \\ = \ln \left( {\frac{{\prod\limits_{1\, \leqslant \,n} n }} {{\prod\limits_{1\, \leqslant \,n} {\left( {n + 1} \right)} }}\;\frac{{\prod\limits_{1\, \leqslant \,n} {\left( {n + 2} \right)} }} {{\prod\limits_{1\, \leqslant \,n} {\left( {n + 1} \right)} }}} \right) = \ln \left( {\frac{{\prod\limits_{1\, \leqslant \,n} n }} {{\prod\limits_{2\, \leqslant \,n} n }}\;\frac{{\prod\limits_{3\, \leqslant \,n} n }} {{\prod\limits_{2\, \leqslant \,n} n }}} \right) = \hfill \\ = \ln \left( {1\;\frac{1} {2}} \right) = \ln \left( {\frac{1} {2}} \right) \hfill \\ \end{gathered} $$

Rewriting the above in more rigorous terms, we have $$ \begin{gathered} \sum\limits_{1\, \leqslant \,n\, \leqslant \,q} {\ln \left( {\frac{{n\left( {n + 2} \right)}} {{\left( {n + 1} \right)^{\,2} }}} \right)} = \ln \left( {\prod\limits_{1\, \leqslant \,n\, \leqslant \,q} {\frac{{n\left( {n + 2} \right)}} {{\left( {n + 1} \right)^{\,2} }}} } \right) = \hfill \\ = \ln \left( {\frac{{\prod\limits_{1\, \leqslant \,n\, \leqslant \,q} n }} {{\prod\limits_{1\, \leqslant \,n\, \leqslant \,q} {\left( {n + 1} \right)} }}\;\frac{{\prod\limits_{1\, \leqslant \,n\, \leqslant \,q} {\left( {n + 2} \right)} }} {{\prod\limits_{1\, \leqslant \,n\, \leqslant \,q} {\left( {n + 1} \right)} }}} \right) = \ln \left( {\frac{{\prod\limits_{1\, \leqslant \,n\, \leqslant \,q} n }} {{\prod\limits_{2\, \leqslant \,n\, \leqslant \,q + 1} n }}\;\frac{{\prod\limits_{3\, \leqslant \,n\, \leqslant \,q + 2} n }} {{\prod\limits_{2\, \leqslant \,n\, \leqslant \,q + 1} n }}} \right) = \hfill \\ = \ln \left( {\frac{1} {{q + 1}}\;\frac{{q + 2}} {2}} \right) \hfill \\ \end{gathered} $$ and therefore $$ \mathop {\lim }\limits_{q\, \to \,\infty } \sum\limits_{1\, \leqslant \,n\, \leqslant \,q} {\ln \left( {\frac{{n\left( {n + 2} \right)}} {{\left( {n + 1} \right)^{\,2} }}} \right)} = \mathop {\lim }\limits_{q\, \to \,\infty } \ln \left( {\frac{1} {2}\;\frac{{q + 2}} {{q + 1}}} \right) = \ln \left( {\frac{1} {2}\mathop {\lim }\limits_{q\, \to \,\infty } \;\frac{{q + 2}} {{q + 1}}} \right) = \ln \left( {\frac{1} {2}} \right) $$

G Cab
  • 35,272
  • 4
    For the sake of the beginner... do note that the argument doesn't make sense literally as written. It is a quick and efficient argument for those with the experience to recognize that, in this particular problem the tails don't pose an issue and everything 'just works'. Those without that experience should take the further steps to turn it into a rigorous argument. –  Jan 04 '17 at 16:52
  • @Hurkyl: If I understand your critics, you are hinting to the problem of preserving convergence when splitting the product, which is in fact a fully right concern, that in this case I "jumped" because absorbed in the "tails" being cut. Thanks for signalling that. – G Cab Jan 04 '17 at 18:52
  • @Hurkyl: now should be better, thanks again for your comment – G Cab Jan 04 '17 at 19:30
7

We can see that $$\frac {n (n+2)}{(n+1)^2} =1-\frac {1}{(n+1)^2} $$ So we have, $$\sum_{n=1}^{\infty} \log \left(1-\frac {1}{(n+1)^2}\right) =\lim_{n \to \infty} \log \left(\left(1-\frac {1}{4}\right)\left(1-\frac {1}{9}\right)\cdots\left(1-\frac {1}{(n+1)^2}\right)\right) =\log \frac {1}{2} $$ Hope it helps.


For why the infinite product is $\frac {1}{2} $, see here.

  • Why does last piece of product (1-1/n^2)? I thought (1-1/(n+1)^2) – Yas Jan 04 '17 at 12:12
  • @Yas, the answer has been edited to change that, but know that the two forms are equivalent. Your sum of $$\sum_{n=1}^\infty \log \left( 1 - \frac{1}{(n+1)^2} \right)$$ can be re-indexed by $m = n+1$ to get $$\sum_{m=2}^\infty \log \left( 1 - \frac{1}{m^2} \right)$$ – CodeLabMaster Jan 05 '17 at 05:41
6

Another overkill. Since: $$ \frac{\sin(\pi x)}{\pi x}=\prod_{n\geq 1}\left(1-\frac{x^2}{n^2}\right)\tag{1} $$ we have: $$ \sum_{n\geq 1}\log\frac{n(n+2)}{(n+1)^2}=\log\prod_{n\geq 2}\left(1-\frac{1}{n^2}\right)=\log\lim_{x\to 1}\frac{\sin(\pi x)}{\pi x(1-x^2)}\stackrel{dH}{=}\color{red}{-\log 2}.\tag{2} $$

Jack D'Aurizio
  • 353,855
5

\begin{align*}\sum_{n=1}^N \ln\left(\frac{n(n+2)}{(n+1)^2}\right)&= \sum_{n=1}^N \left(\ln n + \ln (n+2) -2\ln(n+1)\right)\\&=\ln 1+\ln(N+2)-\ln2-\ln(N+1)\\&=-\ln2+\ln\frac{N+2}{N+1}\\&\xrightarrow{N\to\infty}-\ln2+\ln1=-\ln2.\end{align*} This is however essentially the same solution as that given by Behrouz, just less clever and explicit about the limit.

Carsten S
  • 8,726