6
$\begingroup$

I was trying to learn how to evaluate this expression: $$\sum_{n=0}^{\infty}\frac{(-1)^{n}}{(2n+1)^{3}}$$ And I learned that it can be easily simplified to: $$\frac{1}{64}(\zeta(3,\frac{1}{4})-\zeta(3,\frac{3}{4}))$$ using the Hurwitz Zeta function. From here, I didn't know how to further simplify it to get an answer of: $$\frac{\pi^{3}}{32}$$ On the Mathworld Wolfram page I had seen a simplification like this: $$\zeta(s,\frac{1}{2})=2^{s}\sum_{k=0}^{\infty}\frac{1}{(2k+1)^{s}}=2^{s}(\zeta(s)-\sum_{k=1}^{\infty}\frac{1}{(2k)^{s}})=(2^{s}-1)\zeta(s)$$ But I had no idea how they even did this, or if this is somehow the way that can help to give the simplification to get the answer above.

Edit: I have realized that this is a Dirichlet Beta Function, but now I am wondering if still the original Hurwitz Zeta Function can be simplified without the use of definitions of the Beta Function. (As in, a connection to what I have written above with that simplification)

$\endgroup$
5
  • $\begingroup$ This is a Dirichlet beta function and you obtained the specific value $\displaystyle -1-\beta(3)=-1-\frac{\pi^{3}}{32}$ (because of $2n-1$ instead of $2n+1$) $\endgroup$ Commented Aug 5 at 21:25
  • $\begingroup$ Sorry, I have edited it now to be 2n+1 as 2n-1 was a mistake. I didn't even know about this beta function, and the way which I have obtained the first simplification into a Hurwitz Zeta function was through a different method of simplification which I just did on my own. Then I have one question which is to evaluate this function, do you have to evaluate the integral representation which I have found on wikipedia? $\endgroup$ Commented Aug 5 at 21:35
  • $\begingroup$ I have just realized that if you solve the integral in this case for s=3, you will not even be able to solve the integral as you will get the same series stuck inside ... So how would actually solve for one of the series. $\endgroup$ Commented Aug 5 at 21:53
  • $\begingroup$ $$\sum_{n=0}^{\infty}\frac{(-1)^{n}}{(2n+1)^{2k+1}}=\frac{\pi^{2k+1}}{2^{2k+2}(2k)!}|E_{2k}|$$ encyclopediaofmath.org/wiki/Euler_numbers $\endgroup$
    – Svyatoslav
    Commented Aug 6 at 6:15
  • $\begingroup$ This is interesting because I didn't see this on the dirichlet beta function page, even though they are clearly related. $\endgroup$ Commented Aug 6 at 23:37

1 Answer 1

5
$\begingroup$

This is a Dirichlet beta function and your result may be obtained using this answer.

We may rewrite $\zeta\left(3, \pm \frac {1}4\right)$ using the polygamma function :

$$\tag{1}\psi^{(m)}(z) = (-1)^{m+1} m! \,\zeta(m+1,z)$$

You want $\displaystyle S:=\frac{1}{64}\left(\zeta\left(3,\frac{1}{4}\right)-\zeta\left(3,\frac{3}{4}\right)\right)=\frac{1}{64}\left(-\frac 12\psi^{(2)}\left( \frac {1}4\right)+\frac 12\psi^{(2)}\left( \frac {3}4\right)\right)$

These polygamma values are well known (from Kolbig's paper ) :

\begin{align} \tag{2}\psi^{(2n)}\left(\frac 14\right) &= -2^{2n-1}\left(\pi^{2n+1}|E_{2n}|+2(2n)!(2^{2n+1}-1)\zeta(2n+1)\right)\\ \psi^{(2)}\left(\frac 14\right)&=-2\left(\pi^{3}+2(2)!(2^{3}-1)\zeta(3)\right)\\ \tag{3}\psi^{(2n)}\left(\frac 34\right) &= 2^{2n-1}\left(\pi^{2n+1}|E_{2n}|-2(2n)!(2^{2n+1}-1)\zeta(2n+1)\right)\\ \psi^{(2)}\left(\frac 34\right)&=2\left(\pi^{3}-2(2)!(2^{3}-1)\zeta(3)\right)\\ \end{align}

Subracting the first to the second and dividing by $64$ gives indeed $\displaystyle\frac {\pi^3}{32}$.

We may get the result more directly (cf Kolbig) starting with this polygamma expression : $$\tag{4}\psi^{(n)}(z) = (-1)^{n+1}\, n! \sum_{k=0}^\infty \frac{1}{(z+k)^{n+1}}$$ using the polygamma reflection formula $6.4.7$ (itself derived from the $\Gamma$ function reflection formula $6.1.17$) : $$\tag{5} \psi^{(n)}\left(1-z\right)+(-1)^{n+1}\psi^{(n)}\left(z\right)=(-1)^n\pi \frac {d^n}{dz^n} \cot(\pi z)$$

So that $$\tag{6}\psi^{(2)}\left( \frac {3}4\right)-\psi^{(2)}\left( \frac {1}4\right)=\pi \left.\frac {d^2}{dz^2} \cot(\pi z)\right|_{z=1/4}$$ $$-$$ For a direct derivation use (proofs here ) : $$\tag{7} \pi \cot(\pi z)=\frac 1z+\sum_{n=1}^\infty \frac {2z}{z^2-n^2}=\frac 1z+\sum_{n=1}^\infty \frac 1{z-n}+\frac 1{z+n}$$ and compute the second derivative at $\displaystyle z=\frac 14$.

$\endgroup$
3
  • $\begingroup$ Thank you for the brilliant answer, I didn't realize the answer would be so complicated, but I guess if you memorize only 2 necessary formulas, you could do it by hand (the polygamma-hurwitz relation formula, and the reflection formula). $\endgroup$ Commented Aug 5 at 22:29
  • $\begingroup$ This is not as complicated as you may fear and you may start directly with this expression of the polygamma function $$\psi^{(m)}(z) = (-1)^{m+1}\, m! \sum_{k=0}^\infty \frac{1}{(z+k)^{m+1}}$$ (we simply subtract the positive and negative terms of your series!). Excellent continuation anyway, $\endgroup$ Commented Aug 5 at 22:37
  • $\begingroup$ Yes thank you for your explanation! $\endgroup$ Commented Aug 6 at 23:36

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .