0
$\begingroup$

$$ \pi \sec{\left(\frac{\pi x}{2} \right)} = 4 \sum_{k=0}^{\infty} (-1)^k\frac{2k+1}{(2k+1)^2-x^2} . $$

Preferably avoiding complex analysis or zig-zag numbers. This is my missing link in understanding the 1/(4k+1) series. I can get the similar series for tan (by differentiating log(sin) but have become really stuck with this trying similar approaches and looking for trig identities.

$\endgroup$

1 Answer 1

1
$\begingroup$

Form the sum in the followings:

$S= 4\sum\limits_{k=0}^{\infty} (-1)^k\frac{2k+1}{(2k+1)^2-x^2}=2 \sum\limits_{k=0}^{\infty} \frac{(-1)^k}{2k+1-x}+2\sum\limits_{k=0}^{\infty}\frac{(-1)^k}{2k+1+x}$

Let $k=2m$ if $k$ even and $k=2m+1$ if $k$ odd so we get:

$S=2 \sum\limits_{m=0}^{\infty} \frac{1}{4m+1-x}-2\sum\limits_{k=0}^{\infty}\frac{1}{4m+3-x}+\sum\limits_{m=0}^{\infty} \frac{1}{4m+1+x}-2\sum\limits_{k=0}^{\infty}\frac{1}{4m+3+x}$

Introduce the digamma function: $\psi(1+z)=-\gamma+\sum\limits_{k=1}^\infty\big(\frac{1}{k}-\frac{1}{z+k}\big)$

We have the followings:

$S=\frac{1}{2}\big(-\psi(\frac{x+1}{4})+\psi(\frac{x+3}{4})-\psi(\frac{-x+1}{4})+\psi(\frac{-x+3}{4})\big)$

Using the reflection formula of digamma function:

$\psi(1-z)-\psi(z)=\pi \cot(\pi z)$

We get:

$S=\frac{\pi}{2}\big(\cot(\frac{\pi}{4}(x+1))+\cot(\frac{\pi}{4}(1-x))\big)$

After simplifing the expression using trigonometric laws:

$S=\frac{\pi}{2}\frac{2}{\cos(\frac{x\pi}{2})}=\pi \sec(x\frac{\pi}{2})$

The statement is proved.

$\endgroup$
3
  • $\begingroup$ This will do it and I will learn about digamma function. Also found a related question where putting a=2x is equivalent to my question. $\endgroup$
    – bartfast
    Commented Nov 9, 2018 at 18:44
  • $\begingroup$ I would have a nother solution but there is a heavy integration where I gave it up. The integral: $\int\limits_0^1 \frac{t^x-t^{-x}}{(1+t^2)}dt$. Tomorrow I will deal with it. Perhaps it will help you. $\endgroup$
    – JV.Stalker
    Commented Nov 9, 2018 at 19:09
  • $\begingroup$ thanks. There are parts of the related question's answers I don't yet get so yours is the best answer for me so far. I'm trying to stick to elementary functions and real analysis. I'm also wondering why I can't make a similar approach work to that of deriving the pitan(pix/2) series expansion which seems to come out comparatively easy by differentiating the log of sinx (using the infinite product formula for sinx) $\endgroup$
    – bartfast
    Commented Nov 9, 2018 at 19:37

You must log in to answer this question.

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