MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Good day to everyone.
In my scientific research I've got stuck with a contour integration problem.
I would like to evaluate the following integral: $$I=\int_0^{\infty } \frac{e^{\frac{\alpha -\mathrm i \Omega \beta }{1+\Omega ^2\sigma ^2}}}{\left(1+\Omega ^2\sigma ^2\right)^L\Omega } \, d\Omega \tag{1}$$ Where $\alpha, \beta, \sigma \in \mathbb{R}$ and $\sigma>0$.
First of all I've changed the variable to $x=\sigma\Omega$ and turned the exponent part in partial fractions with $A,B$ depending only on $\alpha, \beta, \sigma$, treating them as constants. $$I=\int_0^{\infty } \frac{e^{\frac{A}{1+\mathrm i x }+\frac{B}{1-\mathrm i x }}}{(1+\mathrm i x )^L(1-\mathrm i x )^Lx } \, dx \tag{1*}$$ Then performed linear fractional transformation (Möbius transformation): $$ \begin{cases} x =\mathrm i\frac{p-1}{p+1} \\ p=-\frac{1-\mathrm i x }{1+\mathrm i x } \end{cases} \tag{2} $$ $$I=\frac{e^{\frac{a}{2}}}{2^{2L-1}}\int _{\gamma }\frac{(1+p)^{2L-1}}{(1-p)p^L}e^{\frac{B}{2}p+\frac{A}{2}\frac{1}{p}}dp\tag{3}$$ where $\gamma$ is and arc in complex p-plane, running counterclockwise from $-1$ to $1$ (is that correct?).
Then using the binomial theorem for $(1+p)^{2L-1}$ I get: $$I=\frac{e^{\frac{a}{2}}}{2^{2L-1}}\int _{\gamma }\frac{\sum_{k=0}^{2L-1} {{2L-1} \choose k}p^k}{(1-p)p^L}e^{\frac{B}{2}p+\frac{A}{2}\frac{1}{p}}dp\tag{4}$$ $$ I=\sum_{k=0}^{2L-1} {{2L-1} \choose k}\frac{e^{\frac{a}{2}}}{2^{2L-1}}\underbrace{\int _{\gamma }\frac{1}{(1-p)p^{L-k}}e^{\frac{B}{2}p+\frac{A}{2}\frac{1}{p}}dp}_{I_1}\tag{4*} $$ So integral $I_1$ is of primary interest.
Form Connections between the Generalized Marcum Q-Function and a class of Hypergeometric Functions by D. Morales-Jimenez, F. J. Lopez-Martinez, E. Martos-Naya, J. F. Paris, A. Lozano I know that $$Q_m(a,b)e^{\frac{A^2+B^2}{2}}=\oint_\Gamma \frac{1}{(1-p)p^{m}}\exp\left(\frac{1}{2}\left({B p+\frac{A}{p}}\right)\right)dp \tag{5}$$ where $Q_m(a,b)$ is the generalized Marcum Q-function and $\Gamma$ is "any closed contour enclosing the singularity at $p = 0$ (in a counterclockwise direction) and no other singularities of the integrand".
So it would be perfect if $\gamma=\Gamma$ then what I have is just a sum of Marcum Q-functions. But it does not.
Ar first I thought about closing the contour in a manner like that
enter image description here
But then realized that I'm not able to cope with integrals alond the real axe (the blue one's).
So, is there any way to cope with this integral? And what about the singularity at $p=1$?
Any hints would be of great help.

share|cite|improve this question
isn't your integral logarithmically divergent at $\Omega=0$? – Carlo Beenakker Sep 26 '13 at 7:01

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.