Help me please with the following integral. I've asked this question before https://math.stackexchange.com/q/140460/30554, but it turns out that it was incorrect question. I should get an asymptotic limit (with $k$ goes to $\infty$, and $q$ fixed) of it. For $q\ge 2, t\ge 0, k \in Z$ $$ \int_0^{\infty}\left(1-\frac{t^2}{2(2k+3)}+\frac{t^4}{2\cdot 4\cdot(2k+3)\cdot(2k+5)}\right)^qdt $$

Thank you.

  • $\begingroup$ Do you expect that this integral converges? At least for me, it is unable to find a way to estimate the behavior of a formula that is always $+\infty$. $\endgroup$ – Sangchul Lee May 5 '12 at 14:53
  • $\begingroup$ "Asymptotic" needs more information. What do you want: $k \to \infty$ with $q$ fixed, for example? $\endgroup$ – GEdgar May 5 '12 at 15:57
  • $\begingroup$ @GEdgar: yes you are right. I've add this information. Thank you. $\endgroup$ – Alex K. May 5 '12 at 18:10


I'm guessing Alex wants hypergeometric $$\begin{align} {}_0F_1(;k+3/2;-t^2/4)&=1-\frac{t^2}{2(2k+3)}+\frac{t^4}{2\cdot 4\cdot(2k+3)\cdot(2k+5)}+\dots \\ &=\Gamma(k+3/2) t^{-k-1/2} 2^{k+1/2} J_{k+1/2}(t) \end{align}$$ inside the parentheses. Then, for example, with $k=10, q=3$ we get $$ \int_0^1 \frac{2147483648\;\Gamma(23/2)^3 \sqrt{2} J_{21/2}(t)^3\,dt}{t^{63/2}} = \frac{2043784548694122266058130317213\pi}{1869392021479944872009840721920} $$


With $f=\Gamma(k+3/2) t^{-k-1/2} 2^{k+1/2} J_{k+1/2}(t)$, Maple 16 has these asymptotic claims as $k \to \infty$: $$\begin{align} \int_0^\infty f\,dt &= \sqrt{\pi}\,k^{1/2} + \frac{3\sqrt{\pi}}{8}\,k^{-1/2} - \frac{7\sqrt{\pi}}{128}\,k^{-3/2}+O(k^{-5/2}) \\ \int_0^\infty f^2\,dt &=\frac{\sqrt{\pi}}{\sqrt{2}}\,k^{1/2}+ \frac{9\sqrt{\pi}}{16\sqrt{2}}\,k^{-1/2} -\frac{31\sqrt{\pi}}{512\sqrt{2}}\,k^{-3/2} + O(k^{-5/2}) \end{align}$$ and did not do $f^3$.

  • $\begingroup$ My original question was with Bessel function. Then I've used Lamb representation to get the hypergeometric fucntion under the integral. But I need asymptotic limit for general case-for any $k$ and $q$. Is it possible to find it? $\endgroup$ – Alex K. May 5 '12 at 18:13

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.