How do you show that

$$\int_{0}^{\infty}\frac{\mathrm{d}t}{t}\,\mathrm{e}^{\cos\left(t\right)}\, \sin\left(\sin\left(t\right)\right) = \frac{\pi}{2}\,\left(\,\mathrm{e} - 1\right)$$

I managed to get the left-hand side to equal the imaginary part of$$I=\int\limits_0^{\infty}\frac {dt}te^{e^{it}}$$But I’m not very sure what to do next. I’m thinking of a substitute $t\mapsto e^{it}$, but I’m not very sure how to evaluate the limit as $t\to\infty$. I also tried contour integration, but I’m not exactly sure what contour to draw.

  • $\begingroup$ Seems like something you'd want to compute using residue theorem, I can't tell you the details right now though. $\endgroup$ – Wojowu Feb 26 '18 at 18:52
  • $\begingroup$ The original integrand is an even function, and the second integrand, viewed as function is meromorphic with a pole at the origin, so you might be able to do with residues, using the boundary of a half-disk indented at the origin as the contour. $\endgroup$ – saulspatz Feb 26 '18 at 18:55
  • $\begingroup$ Yes. Then you have to show that is goes to zero on the semi circle, and compute the residue at $z=0;$ the usual stuff. Oh, and you have to remember to adjust for the fact that you're using a half-circle near the origin. $\endgroup$ – saulspatz Feb 26 '18 at 19:04
  • $\begingroup$ I haven't been able to make the contour I suggested work. I still think an indented contour is the way to go, but it takes a cleverer choice. $\endgroup$ – saulspatz Feb 26 '18 at 19:22
  • $\begingroup$ @Crescendo How are you dealing with the integral on the positive imaginary axis? $\endgroup$ – saulspatz Mar 31 '18 at 17:24

$$e^{\cos t}\sin\sin t=\text{Im}\exp\left(e^{it}\right)=\text{Im}\sum_{n\geq 0}\frac{e^{nit}}{n!}=\sum_{n\geq 1}\frac{\sin(nt)}{n!} $$ and since for any $a>0$ we have $\int_{0}^{+\infty}\frac{\sin(at)}{t}\,dt=\frac{\pi}{2}$ it follows that $$\int_{0}^{+\infty}e^{\cos t}\sin\sin t\frac{dt}{t} = \frac{\pi}{2}\sum_{n\geq 1}\frac{1}{n!}=\frac{\pi}{2}(e-1), $$ pretty simple.

I have a counter-proposal: $$\begin{eqnarray*} \int_{0}^{+\infty}\left(e^{\cos t}\sin\sin t\right)^2\frac{dt}{t^2} &=&\frac{\pi}{2}\sum_{m,n\geq 1}\frac{\min(m,n)}{m!n!}\\&=&-\frac{\pi}{2}I_1(2)+\pi e(e-1)-2\pi e\int_{0}^{1}I_1(2x)e^{-x^2}\,dx. \end{eqnarray*}$$


Here's a way using contour integration. We first consider an indented semi-circular contour in the upper-half of the complex plane.


Since there are no residues inside the contour, we can write down

$$\int\limits_{\epsilon}^{R}dx\, f(x)+\int\limits_{\Gamma_{R}}dz\, f(z)-\int\limits_{\epsilon}^{R}dy\,\frac {e^{e^{-y}}}y+\int\limits_{\gamma_{\epsilon}}dz\, f(z)=0$$


$$f(z)=\frac {e^{e^{iz}}}z$$

It can be shown that as $\epsilon\to0$ and $R\to\infty$, the arc integrals become$$\int\limits_{\Gamma_{R}}dz\, f(z)=\frac {\pi i}2$$$$\int\limits_{\gamma_{\epsilon}}dz\, f(z)=-\frac {e\pi i}2$$

Putting everything together, we see that$$\int\limits_0^{\infty}dx\,\frac {e^{\cos x}}x(\cos\sin x+i\sin\sin x)x-\int\limits_0^{\infty}dy\,\frac {e^{e^{-y}}}y=\frac {\pi i}2(e-1)$$

Take the imaginary part of the first integral and the identity is established.


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.