As a follow up of this nice question I am interested in

$$ S_1=\sum_{m=1}^{\infty}\sum_{n=1}^{m-1}\frac{ 1}{m n\left(m^2-n^2\right)^2} $$

Furthermore, I would be also very grateful for a solution to

$$ S_2=\sum_{m=1}^{\infty}\sum_{n=m+1}^{\infty}\frac{ 1}{m n\left(m^2-n^2\right)^2} $$

Following my answer in the question mentioned above and the numerical experiments of @Vladimir Reshetnikov it's very likely that at least

$$ S_1+S_2 = \frac{a}{b}\pi^6 $$

I think both sums may be evaluated by using partial fraction decomposition and the integral representation of the Polygamma function but I don't know how exactly and I guess there could be a much more efficient route.

  • 1
    $\begingroup$ The inner summation in the first line is summed from $n=0$ to $n-1$. This needs to be adjusted to make sense. As it should also be adjusted in the problem from whence this is first presented. $\endgroup$ – Leucippus May 22 '15 at 14:06
  • $\begingroup$ Sorry for the cross-edits, just making sure the title was covered too $\endgroup$ – abiessu May 22 '15 at 14:13
  • $\begingroup$ no problem. now everthing should be fine... :) $\endgroup$ – tired May 22 '15 at 14:13
  • 2
    $\begingroup$ Where does $m$ start? $\endgroup$ – Alex M. May 22 '15 at 14:15
  • 1
    $\begingroup$ @RenatoFaraone Makes no difference (for $S_1$), since the inner sum is an empty sum for $m = 1$. $\endgroup$ – Daniel Fischer May 22 '15 at 14:28

Clearly, $S_1$=$S_2$ (this can be shown by reversing the order of summation, as was noted above). Using $$ \frac{ 1}{m n\left(m^2-n^2\right)^2}=\frac{ (m+n)^2-(m-n)^2}{4 m^2 n^2\left(m^2-n^2\right)^2} $$ we get $$ S_1=\sum_{m=1}^{\infty}\sum_{n=1}^{m-1}\frac{ 1}{m n\left(m^2-n^2\right)^2}=\frac{1}{4}\sum_{m=1}^{\infty}\sum_{n=1}^{m-1}\frac{ 1}{m^2 n^2\left(m-n\right)^2}-\frac{1}{4}\sum_{m=1}^{\infty}\sum_{n=1}^{m-1}\frac{ 1}{m^2 n^2\left(m+n\right)^2}, $$ and after reversing the order of summation in the first sum $$ S_1=\frac{1}{4}\sum_{n=1}^{\infty}\sum_{m=n+1}^{\infty}\frac{ 1}{m^2 n^2\left(m-n\right)^2}-\frac{1}{4}\sum_{m=1}^{\infty}\sum_{n=1}^{m-1}\frac{ 1}{m^2 n^2\left(m+n\right)^2}=\\ \frac{1}{4}\sum_{n=1}^{\infty}\sum_{m=1}^{\infty}\frac{ 1}{m^2 n^2\left(m+n\right)^2}-\frac{1}{4}\sum_{m=1}^{\infty}\sum_{n=1}^{m-1}\frac{ 1}{m^2 n^2\left(m+n\right)^2}. \qquad\qquad (1) $$

Let's introduce a third sum $$ S_3=\sum_{m=1}^{\infty}\sum_{n=1}^{m-1}\frac{ 1}{m^2 n^2\left(m+n\right)^2} =\sum_{n=1}^{\infty}\sum_{m=n+1}^{\infty}\frac{ 1}{m^2 n^2\left(m+n\right)^2}=\\ \frac{1}{2}\sum_{m=1}^{\infty}\sum_{n=1}^{\infty}\frac{ 1}{m^2 n^2\left(m+n\right)^2}-\frac{1}{8}\zeta(6). $$ Using An Infinite Double Summation $\sum_{k=1}^{\infty} \sum_{n=1}^{\infty} \frac{1}{n^2k^2(n+k)^2}$? we get $$ S_3=\frac{1}{2}\cdot\frac{1}{3}\zeta(6)-\frac{1}{8}\zeta(6)=\frac{1}{24}\zeta(6).\qquad\qquad\qquad (2) $$ From (1) and (2) we get

$$ S_1=\frac{1}{4}\cdot\frac{1}{3}\zeta(6)-\frac{1}{4}\cdot\frac{1}{24}\zeta(6)=\frac{7}{96}\zeta(6)=\frac{7}{96}\frac{\pi^6}{945}=\frac{\pi^6}{12960} $$

  • $\begingroup$ excellent work (+1) , i was failing after your formula one back then....! $\endgroup$ – tired Oct 18 '15 at 10:48

Numerically, I get $$ S_1+S_2 = 0.14836252987273216621 $$ which agrees with $$ \frac{\pi^6}{6480} $$ Also numerically, $$ S_1 = 0.074181264936366083104 \\ S_2 = 0.074181264936366083104 $$ are seemingly equal.

  • $\begingroup$ very well! multiplying this by $8\pi$ and adding $\pi^7/70$gives us the result suggested by @Vladimir Reshetnikov $\endgroup$ – tired May 22 '15 at 15:31
  • $\begingroup$ Do you have an idea where this symmerty come from? I'm not getting my head around it :/ $\endgroup$ – tired May 22 '15 at 15:54
  • 2
    $\begingroup$ Reverse the order of summation in $S_2$ and see what you get. $\endgroup$ – GEdgar May 22 '15 at 16:41
  • $\begingroup$ thanks, oh man today is not my brightest day! $\endgroup$ – tired May 22 '15 at 17:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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