30
$\begingroup$

This is somewhat similar to my previous question: Closed form for $\int_0^1\frac{x^{5/6}}{(1-x)^{1/6}\,(1+2\,x)^{4/3}}\log\left(\frac{1+2x}{x\,(1-x)}\right)\,dx$

Is it possible to find a closed form for this integral? $$Q=\int_0^1\sqrt{\frac{2-x}{(1-x)\,x}}\,\log\left(\frac{(2-x)\,x}{1-x}\right)dx$$

$\endgroup$
  • $\begingroup$ Break it up into three parts, using $\ln\frac a{bc}=\ln a-\ln b-\ln c$. $\endgroup$ – Lucian Nov 14 '13 at 8:09
43
+350
$\begingroup$

First, we transform the integral into a more computable form by using some substitutions.

$$\begin{align*}\displaystyle Q &= \int_0^1\sqrt{\frac{2-x}{(1-x)\,x}}\,\log\left(\frac{(2-x)\,x}{1-x}\right)dx\\ &=\int_0^1 \sqrt{\frac{1+u}{u(1-u)}}\log \left( \frac{(1+u)(1-u)}{u}\right)du \quad \color{blue}{\text{where }u=1-x}\\ &= \int_0^1 \frac{1+u}{\sqrt{u(1-u^2)}}\log \left( \frac{1-u^2}{u}\right)du \\ &= \frac{1}{2}\int_0^1 \frac{1+\sqrt{t}}{t^{\frac{3}{4}}\sqrt{1-t}}\log \left(\frac{1-t}{\sqrt{t}} \right)dt \quad \color{blue}{\text{where }t=u^2} \\ &= \frac{1}{2}\int_0^1 \frac{\log(1-t)}{t^{\frac{3}{4}}\sqrt{1-t}}dt-\frac{1}{4}\int_0^1 \frac{\log(t)}{t^{\frac{3}{4}}\sqrt{1-t}}dt \\ &\quad +\frac{1}{2}\int_0^1 \frac{\log(1-t)}{t^{\frac{1}{4}}\sqrt{1-t}}dt-\frac{1}{4}\int_0^1 \frac{\log(t)}{t^{\frac{1}{4}}\sqrt{1-t}}dt \tag{1} \end{align*}$$

These four integrals can be evaluated by calculating derivatives of beta function in terms of digamma function. For e.g.

$$ \begin{align*} \int_0^1 \frac{\log(1-t)}{t^{\frac{3}{4}}\sqrt{1-t}}dt &= \frac{d}{dz}\left\{ \int_0^1 t^{-\frac{3}{4}}(1-t)^{z-1} \; dt\right\}_{z=\frac{1}{2}}\\&= \frac{d}{dz}\left\{ \frac{\Gamma \left( \frac{1}{4}\right)\Gamma(z)}{\Gamma \left( \frac{1}{4}+z\right)} \right\}_{z=\frac{1}{2}}\\ &= \frac{\Gamma \left( \frac{1}{4}\right)\sqrt{\pi}}{\Gamma \left( \frac{3}{4}\right)}\left\{\psi_0 \left(\frac{1}{2} \right) -\psi_0 \left(\frac{3}{4} \right)\right\}\\ &= \pi^{3/2}\frac{\sqrt{2}}{\Gamma \left( \frac{3}{4}\right)^2}\left\{\log 2-\frac{\pi}{2} \right\} \tag{2} \\ \end{align*} $$ To get the last expression, I used the special values $$ \begin{align*} \psi_0 \left(\frac{3}{4}\right) &= -\gamma +\frac{\pi}{2}-3\log 2 \\ \psi_0 \left(\frac{1}{2}\right) &= -\gamma -2\log 2 \end{align*} $$ Using the same technique, the other three integrals can be evaluated: $$ \begin{align*} \int_0^1 \frac{\log(t)}{t^{\frac{3}{4}}\sqrt{1-t}}dt&= -\pi^{5/2}\frac{\sqrt2}{\Gamma \left( \frac{3}{4}\right)^2}\tag{3} \\ \int_0^1 \frac{\log(t)}{t^{\frac{1}{4}}\sqrt{1-t}}dt&=\frac{(4\pi-16)\Gamma \left( \frac{3}{4}\right)^2}{\sqrt{2\pi}} \tag{4} \\ \int_0^1 \frac{\log(1-t)}{t^{\frac{1}{4}}\sqrt{1-t}}dt&=\frac{2(-8+\pi+2\log 2)\Gamma \left( \frac{3}{4}\right)^2}{\sqrt{2\pi}} \tag{5} \end{align*} $$ Substituting the results of equations $(2),(3),(4)$ and $(5)$ in $(1)$ gives

$$Q=\frac{\Gamma\left(\frac34\right)^{-2}\pi^2\log2-\Gamma\left(\frac34\right)^2(4-2\log2)}{\sqrt{2\,\pi}}$$

$\endgroup$
15
$\begingroup$

$$Q=\frac{\Gamma\left(\frac34\right)^{-2}\pi^2\ln2-\Gamma\left(\frac34\right)^2(4-\ln4)}{\sqrt{2\,\pi}}$$

$\endgroup$
  • 19
    $\begingroup$ Cleo, can you enlighten us as to how you got this result? $\endgroup$ – Ron Gordon Nov 14 '13 at 21:45
  • 9
    $\begingroup$ (-1) downvoted because this only throw us an answer w/o giving us any hint/ref how to get it. $\endgroup$ – achille hui Nov 19 '13 at 8:27
  • 14
    $\begingroup$ Do you downvote Fermat's post, which was made 350 years ago, too? $\endgroup$ – Math.StackExchange Nov 19 '13 at 8:40
  • 24
    $\begingroup$ @AranKomatsuzaki I won't downvote Fermat's comment on a margin because it is a comment. Math.SE is an Q & A site, being an answer, in particular one with a big bonus attached, we will expect one with more details included. $\endgroup$ – achille hui Nov 19 '13 at 8:45
  • 7
    $\begingroup$ @RonGordon It happens frequently over here. Sometimes ago, I make a lot of calculations to derive an answer and I got one upVote from the OP. However, somebody else wrote four lines of a 'philosophical speech' but never wrote any answer. He or she just said something like 'it's trivial' or so. It looks like there are 'circle of friends' over here. $\endgroup$ – Felix Marin Nov 24 '13 at 3:15

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.