4
$\begingroup$

I entered this integration problem to Mathematica Online Integrator an got a solution I would never have been able to find manually.

$$\int\root 3 \of{\cos(x)^2}\,dx=\frac{(-3\cos(x)\root 3 \of{\cos(x)^2}*Hypergeometric2F1[1/2, 5/6, 11/6, \cos(x)^2]*\sin(x))}{5\sqrt{\sin(x)^2}}$$

I would like to see/learn how the trick works to solve/reproduce this the manual way.

$\endgroup$
  • $\begingroup$ It's really ugly answer to me. $\endgroup$ – Tunk-Fey May 27 '14 at 16:37
  • 1
    $\begingroup$ Use the binomial series and then switch the order of summation and integration. $\endgroup$ – Lucian May 27 '14 at 19:26
  • $\begingroup$ I also was disappointed that such a "nice" integral results in such an "ugly" answer. $\endgroup$ – Wolfgang Tintemann May 28 '14 at 15:55
4
$\begingroup$

Expanding the above comment, we can write (say, for $x\in\left[0,\frac{\pi}{2}\right]$) \begin{align} \int \cos^{r}x\,dx&=-\int \cos^r x\left(1-\cos^2 x\right)^{-\frac12}\,d(\cos x)=\\ &=-\int \sum_{k=0}^{\infty}\frac{\Gamma(k+\frac12)}{k!\,\Gamma(\frac12)}\cos^{r+2k}x\,d(\cos x)=\\ &=-\sum_{k=0}^{\infty}\frac{\Gamma(k+\frac12)}{k!\,\Gamma(\frac12)}\frac{\cos^{r+1+2k}x}{r+1+2k}=\\ &=-\frac{\cos^{r+1}x}{r+1}\sum_{k=0}^{\infty}\frac{(\frac12)_k(\frac{r+1}{2})_k}{k!\,(\frac{r+3}{2})_k}\left(\cos^2 x\right)^k=\\ &=-\frac{\cos^{r+1}x}{r+1}{}_2F_1\left(\frac12,\frac{r+1}{2};\frac{r+3}{2};\cos^2 x\right),\tag{1} \end{align} where $(a)_k=\frac{\Gamma(a+k)}{\Gamma(a)}$ denotes the Pochhammer symbol. It remains to set $r=\frac23$ in the last formula.

$\endgroup$
  • $\begingroup$ Thats exactly the answer I was looking for. The mystery of solution is now solved. I have a copyright-oriented question : is it allowed to use this result/way of solution in an own perhaps copyright-protected script publishing of course the source and author ? $\endgroup$ – Wolfgang Tintemann May 28 '14 at 15:47
  • $\begingroup$ Please feel free to use this without any referencing. $\endgroup$ – Start wearing purple May 28 '14 at 21:43
  • $\begingroup$ If I set $r=0$ in the above equation there results a funny formula. $\endgroup$ – Wolfgang Tintemann May 29 '14 at 15:49
  • $\begingroup$ I noticed the following. $$\int_{-\frac \pi 2}^{\frac \pi 2} \root 3 \of{\cos^2(x)}\,dx$$ is greater zero. But if I try to calculate with the solution given I get zero ( I used WolframAlpha ). This seems not to be an error using the CAS as $$\cos(x)=0$$ for $$\pm \frac \pi 2$$. $\endgroup$ – Wolfgang Tintemann May 29 '14 at 18:26
  • $\begingroup$ @WolfgangTintemann Good point. The seeming contradiction is related to that for $x\in(-\pi/2,0)$ we should replace (at the very first step) $\sin x$ by $-\sqrt{1-\cos^2x}$ (i.e. there is a sign difference). If you want to have an antiderivative valid for $ x\in(-\pi/2,\pi/2)$, in addition to modifying the sign one should add a constant to one of the two expressions ensuring that they match at $x=0$. $\endgroup$ – Start wearing purple May 29 '14 at 22:33

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.