4
$\begingroup$

Prove that $$\int_0^{1/e} \frac{\mathrm{dx}}{\sqrt{(\ln x)^2-1}}=K_{0}(1)$$ where $K_{n}(x)$ is the modified bessel function of the second kind.
Some hints/suggestions?
Thanks.

$\endgroup$
  • $\begingroup$ Have you tried changing variables to $u=\ln x$, and then going by parts? $\endgroup$ – user54147 Dec 26 '12 at 19:08
  • $\begingroup$ @LevLivnev: I thought of $x=e^u$. I'll go that way right now. Thanks. $\endgroup$ – user 1357113 Dec 26 '12 at 19:09
3
$\begingroup$

Let $t = \ln(x)$. Then the integral becomes: $$ \int_{-\infty}^{-1} \frac{1}{\sqrt{t^2-1}} \mathrm{e}^{t} \mathrm{d} t \stackrel{u=-t}{=} \int_{1}^\infty \frac{\exp(-u)}{\sqrt{u^2-1}} \mathrm{d} u $$ Now compare this to eq. 10.3.28 of the DLMF handbook of special functions. $$ \int_{1}^\infty \frac{\exp(-u)}{\sqrt{u^2-1}} \mathrm{d} u = \left.\frac{\Gamma\left(\nu+1/2\right)}{ \sqrt{\pi} \left(z/2\right)^{\nu}} K_\nu(z) \right|_{\nu=0, z=1} = K_0(1) $$

$\endgroup$
  • $\begingroup$ just brilliant! $\endgroup$ – user 1357113 Dec 26 '12 at 19:21
2
$\begingroup$

Here is a related problem. Use the change of variable $x=e^{-(t+1)}$ and you will find that the new integrand behaves like $c\,t^{-1/2}$ as $t\to 0$.

Edit: Now, to evaluate the integral, we use the change of variables $x=e^{-(t+1)}$

$$ \int_0^{1/e} \frac{\mathrm{dx}}{\sqrt{(\ln x)^2-1}}=\frac{1}{e}\int _{0}^{\infty }\!{\frac {{{\rm e}^{-t}}}{\sqrt {t}\sqrt {t+2}} }{dt} . $$

We will consider the more general integral

$$ \int _{0}^{\infty }\!{\frac {{{\rm e}^{-st}}}{\sqrt {t}\sqrt {t+2}} }{dt}. $$

The above integral is nothing but the Laplace transform of the function $\frac{1}{\sqrt{t^2+2t}}$ which is given by

$$e^sK_0(s).$$

So, we have

$$ \frac{1}{e}\int _{0}^{\infty }\!{\frac {{{\rm e}^{-t}}}{\sqrt {t}\sqrt {t+2}} }{dt}=\frac{1}{e}\lim_{s\to 1} e^sK_{0}(s)=K_{0}(1). $$

$\endgroup$
  • 1
    $\begingroup$ I'm not sure I understand your point. Does it prove that $\int_0^{1/e} \frac{\mathrm{dx}}{\sqrt{(\ln x)^2-1}}=K_{0}(1)$? $\endgroup$ – user 1357113 Dec 26 '12 at 19:18
  • 1
    $\begingroup$ This answer does not seem to address the OP's question at all. $\endgroup$ – Sasha Dec 26 '12 at 19:22
  • $\begingroup$ @Chris'ssister: I added more materials. $\endgroup$ – Mhenni Benghorbal Dec 27 '12 at 1:55

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.