I am trying to evalute this integral: $$I = \int_0^\infty \frac{1}{x^p(2+x)} \quad p \in \mathbb{R}$$
For the integral to be convergent, it must be: $0<p<1$, right?
I use a pacman/keyhole contour as a contour and I call it C.
$$f(z) = \frac{1}{z^p(2+z)}$$ We have only one pole at $z=-2$: $$\text{Res}_{z=-2} f(z) = \frac{1}{(-2)^p}$$
I use the residue theorem. I know that the: $$\oint_C f(z) dz = 2\pi i \frac{1}{(-2)^p}$$
and: $$\oint_C f(z) dz = \int_0^\infty \frac{1}{x^p(2+x)} - e^{-2\pi i p} \int_0^\infty \frac{1}{x^p(2+x)} = I(1-e^{-2\pi i p}) = 2\pi i \frac{1}{(-2)^p}$$
I get that:
$$I = 2\pi i \frac{(-2)^{-p}}{(1-e^{-2\pi i p})} = 2\pi i\frac{(-2)^{-p}}{(e^{\pi i p}-e^{- i \pi p})e^{-i \pi p}} = \frac{\pi (-2)^{-p}}{\sin(\pi p)e^{-i \pi p}}$$
I am afraid there is some mistake. Thank you for your help.
EDIT:
Noting that: $$e^{-\pi i p} = (e^{i\pi})^{-p} = (-1)^{-p} $$ $$I = \frac{\pi}{\sin(\pi p)} \left(\frac{-2}{-1}\right)^{-p} = \frac{\pi 2^{-p}}{\sin(\pi p)}$$
$$\frac{1}{x^p(x+2)}=\frac{x^{1-a}}{x^{\lfloor p\rfloor}(x+2)}\frac1x$$
Then you get a typical integral of the Mellin type:
$$\mathcal{M}f(b)=\int^\infty_0 f(x)x^b\frac{dx}{x}$$ with $b=1-a>0$
– Mittens Feb 11 '25 at 17:10