Consider that I am some height $r$ above a cylinder (measured from the centre) that has its axis in the direction of $z$. I wish to integrate a 2D Gaussian (of SD $\sigma$) along some part of its length and have the following double integral: $$ \int_0^{2\pi} \int_0^L \exp\left[-(a^2 + r^2 + z^2 - 2ar\cos\phi)/2\sigma^2\right] \mathrm{d}z\, \mathrm{d}\phi, $$ where $a$ is the radius of the cylinder. I can then rearrange this to form: $$ \int_0^{2\pi} \exp \left[ -(a^2 + r^2 - 2ar\cos\phi) / 2\sigma^2 \right] \mathrm{d}\phi \int_0^L \exp \left[ -z^2 / 2\sigma^2 \right] \mathrm{d}z. $$ The second integral gives $$ \DeclareMathOperator{\erf}{erf} \sqrt{\frac{\pi}{2}} \sigma \erf \left( \frac{L}{\sqrt{2} \sigma} \right), $$ but I can't see any way of doing the first integral. Wolfram|Alpha also seems to agree. Have I missed a trick, or is this indeed not do-able with standard functions? Is it possible that a solution could be found if I parameterised my problem differently?


Use the fact that

$$\int_0^{2 \pi} d\phi \, e^{x \cos{\phi}} = 2 \pi I_0(x) $$

where $I_0$ is the modified Bessel function of the first kind of zeroth order.

  • $\begingroup$ Great, thanks a lot! $\endgroup$ – Student's Teapot Oct 23 '15 at 14:27

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.