The following integral $$\int\limits_0^\infty \frac{\cos{\left(\frac{1}{2}\sqrt{3}s\right)}}{\sqrt{\cosh{s}-\cos{\theta}}}\,ds$$ can be found in the paper

Tevian Dray and Gerard 't Hooft, The gravitational shock wave of a massless particle, Nuclear Physics B 253 (1985) 173--188, doi:10.1016/0550-3213(85)90525-5.

They write that they "have not attempted to perform the integration explicitly". Was this integral ever calculated explicitly?

  • $\begingroup$ Do you mean to find closed form or just numeric approximation? $\endgroup$
    – joro
    Feb 12, 2016 at 9:58
  • 3
    $\begingroup$ Wolfram alpha and sage don't seem to know how to compute this integral exactly. $\endgroup$
    – Ben McKay
    Feb 12, 2016 at 10:52
  • $\begingroup$ @joro I mean to find a closed form, if this is possible. $\endgroup$ Feb 12, 2016 at 12:49
  • 1
    $\begingroup$ See Conical Function mathworld.wolfram.com/ConicalFunction.html $\endgroup$ Feb 13, 2016 at 8:46

1 Answer 1


A closed form exists in Gradshteyn and Ryzhik, 8.842.1.

As a comment to this answer:

Referring to the Gradshteyn and Ryzhik, the following result is given in http://arxiv.org/abs/hep-th/9408169 (On Gravitational Shock Waves in Curved Spacetimes, by K. Sfetsos): $$\int \limits_0^\infty \frac{\cos{(\sqrt{c-1/4}\,s)}}{\sqrt{\cosh{s}-\cos{\theta}}}\,ds=\frac{\pi}{\sqrt{2}\cosh{(\sqrt{c-1/4}\,\pi)}}F\left(\frac{1}{2}-i\sqrt{c-1/4},\frac{1}{2}+i\sqrt{c-1/4},1,\cos^2{\frac{\theta}{2}}\right)=\frac{\pi}{\sqrt{2}\cosh{(\sqrt{c-1/4}\,\pi)}}P_{-1/2+i\sqrt{c-1/4}}(-\cos{\theta}). $$

  • 2
    $\begingroup$ Would you mind sharing it with those of us who don't have access to the book? In Google Book's preview the page with with the formula is unfortunately omitted. $\endgroup$
    – eins6180
    Feb 13, 2016 at 6:22
  • $\begingroup$ See also mathworld.wolfram.com/ConicalFunction.html $\endgroup$ Feb 13, 2016 at 8:45
  • $\begingroup$ If this is in terms of the Legendre function of the first kind (en.wikipedia.org/wiki/Legendre_function), then it's like saying "Erf is the solution to the integral...". Although one can (as usual) reduce to a certain hypergeometric function, which I guess as explicit as one can usually get in these cases. $\endgroup$
    – David Roberts
    Feb 13, 2016 at 8:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.