13
$\begingroup$

My nephew asked me this, so I suggested him to sign up here. But anyways this question I was trying to solve myself. I got part of the solution. Let me know the rest.

$(1)$ Solve the integral defined as

$\displaystyle{S_k = (-1)^k \int_0^1 (\log(\sin \pi x))^k dx}$ and show

$(2)$ $\displaystyle{S_k = \frac{(-1)^k}{\sqrt\pi 2^k} \frac{d^k}{d\alpha^k} \frac{\Gamma(\alpha+\frac{1}{2})}{\Gamma(\alpha+1)}}$ with $\alpha=0$.

$(3)$ Show that $\displaystyle{S_4 = \frac{19 \pi^4}{240}+\frac{1}{2} \pi^2 \log^2 2 + \log^4 2 + 6 \log 2 \, \zeta(3)}$

$(4)$ Show the following: $\displaystyle{\int_0^1 \log \log \left(\frac{1}{x}\right) \frac{dx}{1+x^2} = \frac{\pi}{2}\log \left(\sqrt{2\pi} \Gamma\left(\frac{3}{4}\right) / \Gamma\left(\frac{1}{4}\right)\right)}$

For $(3)$, if I substitute $y=\pi x$, I can transform this to a well known log-sine function as described here at Wolfram, and so showing $\displaystyle{S_4 = \frac{19 \pi^4}{240}+\frac{1}{2} \pi^2 \log^2 2 + \log^4 2 + 6 \log 2 \, \zeta(3)}$ is not hard.

I would like suggestions to read or partial solutions (not complete solutions).

(NOTE: I also noted that someone have asked a similar question for $k=2$ here)

$\endgroup$
  • $\begingroup$ So what the difference between problems (1) and (2)? Problem (1) supposes that $S_k$ have another and more better deccription then in (2)? $\endgroup$ – Norbert Mar 17 '12 at 21:45
  • 1
    $\begingroup$ Looks like $(1)$ and $(2)$ are the same. $\endgroup$ – Kirthi Raman Mar 17 '12 at 23:22
  • $\begingroup$ I have understood the two solutions and can take from there. But I wonder how $(4)$ was related to this. Or should I have asked that as separate question? $\endgroup$ – Kirthi Raman Mar 17 '12 at 23:35
  • $\begingroup$ @KirthiRaman: I think (4) deserves its own separate question. $\endgroup$ – Eric Naslund Mar 17 '12 at 23:48
  • $\begingroup$ Is there a way to accept both the answers. (I feel bad because both are good for me) $\endgroup$ – Kirthi Raman Mar 18 '12 at 0:56
13
$\begingroup$

As for the the problem (2) we have $$ \begin{align} S_k&=(-1)^k \int\limits_0^1\log^k(\sin\pi x)dx\\ &=\frac{(-1)^k}{\pi} \int\limits_0^\pi\log^k(\sin y)dy\\ &=\frac{(-1)^k}{\pi} \int\limits_0^{\pi/2}\log^k(\sin y)dy+ \frac{(-1)^k}{\pi} \int\limits_{\pi/2}^\pi\log^k(\sin y)dy\\ &=\frac{(-1)^k}{\pi} \int\limits_0^{\pi/2}\log^k(\sin y)dy+ \frac{(-1)^k}{\pi} \int\limits_{0}^{\pi/2}\log^k(\sin y)dy\\ &=\frac{2(-1)^k}{\pi} \int\limits_0^{\pi/2}\log^k(\sin y)dy\\ &=\frac{2(-1)^k}{\pi} \int\limits_0^{1}\frac{\log^k(t)}{\sqrt{1-t^2}}dt \end{align} $$ Now following Sasha's idea consider integral $$ I(\alpha)= \int\limits_0^1\frac{t^{2\alpha}}{\sqrt{1-t^2}}dt= \frac{1}{2}B\left(\frac{1}{2},\frac{2\alpha+1}{2}\right)= \frac{\sqrt{\pi}}{2}\frac{\Gamma\left(\frac{2\alpha+1}{2}\right)}{\Gamma\left(\frac{2\alpha+2}{2}\right)} $$ Note that $$ I^{(k)}(\alpha)=\int\limits_0^1\frac{t^{2\alpha}2^k\log^k(t)}{\sqrt{1-t^2}}dt $$ so $$ S_k= \frac{2(-1)^k}{\pi} \int\limits_0^{1}\frac{\log^k(t)}{\sqrt{1-t^2}}dt= \frac{2(-1)^k}{\pi} \frac{1}{2^k}I^{(k)}(0)= \frac{(-1)^k}{2^k\sqrt{\pi}}\frac{d^k}{d\alpha^k}\left(\frac{\Gamma\left(\alpha+\frac{1}{2}\right)}{\Gamma\left(\alpha+1\right)}\right)_{\alpha=0} $$

$\endgroup$
17
$\begingroup$

Problem (2): Use generating series. Consider the function $$f(z)=\sum_{k=0}^{\infty}S_{k}\frac{z^{k}}{k!}.$$ Then $$f(z)=\int_0^1 \left(\sum_{k=0}^\infty (-1)^k \log^k(\sin(\pi x)) z^k \right)dx=\int_{0}^{1}\frac{1}{\sin\left(\pi x\right)^{z}}dx=\frac{\Gamma\left(\frac{1-z}{2}\right)}{\sqrt{\pi}\Gamma\left(1-\frac{z}{2}\right)}.$$ The last equality follows from an identity of the Beta Function and then applying the Duplication Formula. From here, differentiating and plugging in $z=0$ allows you to conclude (2).

(You said you wanted few details)

$\endgroup$
  • $\begingroup$ Sorry Eric, your answer was useful too, but I can only accept one answer. $\endgroup$ – Kirthi Raman Mar 18 '12 at 1:04
  • $\begingroup$ @Eric Very nice work! I've recently got this book about generating functions (it had a long alltogethername ending in "ilogy" but I don't remember the name now). The theory is very interesting! $\endgroup$ – Pedro Tamaroff Mar 18 '12 at 1:25
  • $\begingroup$ That's "Generatingfunctionology"! I also like "A=B". $\endgroup$ – marty cohen Mar 18 '12 at 4:52
  • 3
    $\begingroup$ One can download generatingfunctionology and A=B. $\endgroup$ – Did Mar 18 '12 at 9:56
  • $\begingroup$ @Didier That one! $\endgroup$ – Pedro Tamaroff Mar 18 '12 at 15:53

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.