Taken from the post: The Integral that Stumped Feynman?

I want to know if the integral:

$$\int_0^{2\pi} \exp\left(\frac{7+5 \cos x}{10+6\cos x}\right) \cos \left( \frac{\sin x}{10+6 \cos x} \right) dx = 2\pi e^{2/3}$$

can be evaluated using strictly real methods.

I've tried series of $e^x$ and $\cos x$ but to no avail. I tried differentiating under the integral, but nothing seemed to come out of it. Is there any wizardry that can conjure up this answer without complex analysis.

  • $\begingroup$ Would Weierstrass substitution help? $\endgroup$ – Henry Lee Sep 10 '18 at 11:34
  • $\begingroup$ Just gave it a try, didn't seem to get anywhere. $\endgroup$ – Tom Himler Sep 10 '18 at 14:53
  • $\begingroup$ Hmmm. Is there no way you could use complex analysis rather than strictly real? $\endgroup$ – Henry Lee Sep 10 '18 at 14:54
  • $\begingroup$ Since exp(a+b)=exp(a)exp(b) then the equality is equivalent to something like: $\int$ exponential() cosinus()=$2\pi$ $\endgroup$ – FDP Sep 10 '18 at 15:47
  • 1
    $\begingroup$ Does this solution not count? $\endgroup$ – Caddyshack Sep 11 '18 at 19:52

Start with $\tan \left(\frac{x}{2} \right)$$=2\tan \left( \frac{t}{2} \right)$ $$I=2\int_0^{\pi} \exp\left(\frac{7+5 \cos x}{10+6\cos x}\right) \cos \left( \frac{\sin x}{10+6 \cos x} \right) dx=8e^{5/8}\int_0^{\pi}\exp\left(\frac{\cos t}{8}\right)\cos\left(\frac{\sin t}{8}\right)\frac{dt}{5-3\cos t} $$ Using $$ \sum_{n=1}^{\infty} a^{n} \cos(nx) = \frac12\left(\frac{1-a^{2}}{1-2 a \cos x + a^{2}}-1\right)$$ we can rewrite $$\frac{1}{5-3\cos x} =\frac14 +\frac12 \sum_{n=1}^\infty \frac{1}{3^n} \cos (nx) $$ thus we have $$I=2e^{5/8} \int_0^\pi \exp\left(\frac{\cos t}{8}\right)\cos\left(\frac{\sin t}{8}\right) dt +4e^{5/8} \sum_{n=1}^\infty \frac{1}{3^n} \int_0^\pi \exp\left(\frac{\cos t}{8}\right)\cos\left(\frac{\sin t}{8}\right) \cos(nt )dt $$ $$=2\pi e^{5/8}+4e^{5/8}\sum_{n=1}^\infty \frac{1}{3^n} I(n)$$ I dont know how to evaluate $I(n)$, but maybe someone can help. $$I(0)=\pi,I(1)=\frac{\pi}{2^4}, I(2)=\frac{\pi}{2^8}, I(3)=\frac{\pi}{3\cdot 2^{11}}, I(4)=\frac{\pi}{3\cdot2^{16}},I(5)=\frac{\pi}{3\cdot 5 \cdot 2^{19}}$$$$ I(6)=\frac{\pi}{3^2\cdot5\cdot 2^{23}}, I(10)=\frac{\pi}{3^4\cdot5^2\cdot 7 \cdot 2^{39}}, I(20)=\frac{\pi}{3^8\cdot5^4 \cdot 7^2\cdot11\cdot 13\cdot 17\cdot 19\cdot 2^{79}}$$

Edit: As seen here: https://math.stackexchange.com/a/2913057/515527 $ I(n) =\frac{\pi} {2^{3n+1}n!}$, plugging this into the sum and using the series for $e^x$ will give the result immediately..

Another approach to evaluate $I(n)$ is to use that $$\exp\left(\frac{\cos t}{8}\right)\cos\left(\frac{\sin t}{8}\right)=\sum_{n=0}^{\infty} \frac{\cos(nt)}{8^nn!}$$ Since $$\int_0^\pi \cos(nx) \cos(mx) dx=\begin{cases} \frac{\pi} {2} & n=m \\ 0 & n \neq m\end{cases}$$ We get that $I(n) =\frac{\pi} {2} \frac{1} {8^n n!} $ and the result follows.

  • $\begingroup$ We might conjecture that $I(n) = \pi/2^{4n}$ $\endgroup$ – user370967 Sep 11 '18 at 8:40
  • 3
    $\begingroup$ Well, this way we get $I=\frac{98}{47}\pi e^{5/8}$, Unfortunately $ \frac{98}{47}\pi e^{5/8}-2\pi e^{2/3}= 0.000073...$ :( $\endgroup$ – Zacky Sep 11 '18 at 8:54
  • $\begingroup$ Yes, the patterns unfortunately breaks down. $\endgroup$ – user370967 Sep 11 '18 at 10:07
  • $\begingroup$ By the way, I'm not sure you can prove the series identity used by real methods. $\endgroup$ – FDP Sep 11 '18 at 16:31
  • $\begingroup$ look here: math.stackexchange.com/questions/685278/… $\endgroup$ – Zacky Sep 11 '18 at 18:40

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.