The sum series exercise started as: $$\sum_{n=1}^\infty\frac{\ln\left(\frac{n^n}{\left(n+1\right)^n}\right)}{n(n+1)} = \sum_{n=1}^\infty \frac{n\ln\frac{n}{n+1}}{n(n+1)} = \sum_{n=1}^\infty \frac{\ln(n) - \ln(n+1)}{n+1}$$ Looking into calculating it. A quick hint will help me out!
-
Please what is the question? – Olivier Oloa Dec 31 '18 at 12:02
-
Oh, sorry. I am looking into calculating it. – Theodossis Papadopoulos Dec 31 '18 at 12:05
-
A numerical approximation or a closed form? – Olivier Oloa Dec 31 '18 at 12:06
-
Whatever of both is going to help me. Even giving me hints like "steps.. and a telescope series" – Theodossis Papadopoulos Dec 31 '18 at 12:12
-
Use the comparison test – Dr. Sonnhard Graubner Dec 31 '18 at 12:22
-
1Convergence of the series is direct; the value of its sum, though, is not obvious. What makes you think it can be computed? – Did Dec 31 '18 at 12:23
-
It can be computed, because its on my exercise book of my University. – Theodossis Papadopoulos Dec 31 '18 at 12:35
-
1Im studying some quick exercises that our Series teacher gave us for Christmas (Im in Kapodistrian University of Athens at Mathematics department). The exercise just says: Calculate: and the sum of the first part my question above. – Theodossis Papadopoulos Dec 31 '18 at 12:44
-
Perhaps try to view this sum as a reimann sum of an integral (Probably improper ) – Sar Dec 31 '18 at 12:51
-
@Sar No, he just want us with tricks and stuff to end to a telescope series. Olivier gave a perfect answer that Im digging into to find out a faster solution. – Theodossis Papadopoulos Dec 31 '18 at 12:58
-
"Calculate" seems to refer to expressing this series as the value of a rather unusual function defined as... a series rather closely related to the initial ones. One can have doubts about the usefulness of the exercise and the understanding it helps to acquire. – Did Dec 31 '18 at 13:05
-
@Did exactly this – Theodossis Papadopoulos Dec 31 '18 at 13:11
1 Answers
As $n \to \infty$, the general term of the series satisfies $$ \frac{\ln\left(\frac{n^n}{(n+1)^n}\right)}{n(n+1)}=-\frac{1}{n+1}\ln{\small{\left(1+\frac1n\right)}}\sim -\frac1{n^2} $$ giving the convergence of the series.
From $$ \sum_{n=1}^\infty\frac{\ln\left(\frac{n^n}{(n+1)^n}\right)}{n(n+1)}=\sum_{n=1}^\infty\frac{\ln n-\ln(n+1)}{n+1} $$ one may use Theorem 2 (16) to get a closed form in terms of poly-Stieltjes constants:
$$ \sum_{n=1}^\infty\frac{\ln\left(\frac{n^n}{(n+1)^n}\right)}{n(n+1)}=\gamma_1(0,1)-\gamma_1 \tag1 $$
where $$ \gamma_1(a,b) = \lim_{N\to+\infty}\left(\sum_{n=1}^N \frac{\log (n+a)}{n+b}-\frac{\log^2 \!N}2\right) $$ and $\gamma_1=\gamma_1(1,1)$ is an ordinary Stieltjes constant.
By using Mathematica, one gets
$$ \sum_{n=1}^\infty\frac{\ln\left(\frac{n^n}{(n+1)^n}\right)}{n(n+1)}=-0.7885205660\cdots. \tag2 $$

- 120,989
-
1
-
-
Please, I'm not sure I get your point. Do you mean you can obtain a closed form with a telescoping series? – Olivier Oloa Dec 31 '18 at 14:24