115

I've been looking at

$$\int\limits_0^\infty {\frac{{{x^n}}}{{1 + {x^m}}}dx }$$

It seems that it always evaluates in terms of $\sin X$ and $\pi$, where $X$ is to be determined. For example:

$$\displaystyle \int\limits_0^\infty {\frac{{{x^1}}}{{1 + {x^3}}}dx = } \frac{\pi }{3}\frac{1}{{\sin \frac{\pi }{3}}} = \frac{{2\pi }}{{3\sqrt 3 }}$$

$$\int\limits_0^\infty {\frac{{{x^1}}}{{1 + {x^4}}}dx = } \frac{\pi }{4}$$

$$\int\limits_0^\infty {\frac{{{x^2}}}{{1 + {x^5}}}dx = } \frac{\pi }{5}\frac{1}{{\sin \frac{{2\pi }}{5}}}$$

So I guess there must be a closed form - the use of $\Gamma(x)\Gamma(1-x)$ first comess to my mind because of the $\dfrac{{\pi x}}{{\sin \pi x}}$ appearing. Note that the arguments are always the ratio of the exponents, like $\dfrac{1}{4}$, $\dfrac{1}{3}$ and $\dfrac{2}{5}$. Is there any way of finding it? I'll work on it and update with any ideas.


UPDATE:

The integral reduces to finding

$$\int\limits_{ - \infty }^\infty {\frac{{{e^{a t}}}}{{{e^t} + 1}}dt} $$

With $a =\dfrac{n+1}{m}$ which converges only if

$$0 < a < 1$$

Using series I find the solution is

$$\sum\limits_{k = - \infty }^\infty {\frac{{{{\left( { - 1} \right)}^k}}}{{a + k}}} $$

Can this be put it terms of the Digamma Function or something of the sort?

Pedro
  • 122,002

11 Answers11

81

I would like to make a supplementary calculation on BR's answer.

Let us first assume that $0 < \mu < \nu$ so that the integral $$ \int_{0}^{\infty} \frac{x^{\mu-1}}{1+x^{\nu}} \; dx $$ converges absolutely. By the substitution $x = \tan^{2/\nu} \theta$, we have $$ \frac{dx}{1+x^{\nu}} = \frac{2}{\nu} \tan^{(2/\nu)-1} \theta \; d\theta. $$ Thus $$ \begin{align*} \int_{0}^{\infty} \frac{x^{\mu-1}}{1+x^{\nu}} \; dx & = \int_{0}^{\frac{\pi}{2}} \frac{2}{\nu} \tan^{\frac{2\mu}{\nu}-1} \theta \; d\theta \\ & = \frac{1}{\nu} \beta \left( \frac{\mu}{\nu}, 1 - \frac{\mu}{\nu} \right) \\ & = \frac{1}{\nu} \Gamma \left( \frac{\mu}{\nu} \right) \Gamma \left( 1 - \frac{\mu}{\nu} \right) \\ & = \frac{\pi}{\nu} \csc \left( \frac{\pi \mu}{\nu} \right), \end{align*} $$ where the last equality follows from Euler reflexion formula.

Sangchul Lee
  • 167,468
39

The general formula (for $m > n+1$ and $n \ge 0$) is $\frac{\pi}{m} \csc\left(\frac{\pi (n+1)}{m}\right)$. IIRC the usual method involves a wedge-shaped contour of angle $2 \pi/m$.

EDIT: Consider $\oint_\Gamma f(z)\ dz$ where $f(z) = \frac{z^n}{1+z^m}$ (using the principal branch if $m$ or $n$ is a non-integer) and $\Gamma$ is the closed contour below:

enter image description here

$\Gamma_1$ goes to the right along the real axis from $\epsilon$ to $R$, so $\int_{\Gamma_1} f(z)\ dz = \int_\epsilon^R \frac{x^n\ dx}{1+x^m}$. $\Gamma_3$ comes in along the ray at angle $2 \pi/m$. Since $e^{(2 \pi i/m) m} = 1$, $\int_{\Gamma_3} f(z)\ dz = - e^{2 \pi i (n+1)/m} \int_{\Gamma_1} f(z)\ dz$. $\Gamma_2$ is a circular arc at distance $R$ from the origin. Since $m > n+1$, the integral over it goes to $0$ as $R \to \infty$. Similarly, the integral over the small circular arc at distance $\epsilon$ goes to $0$ as $\epsilon \to 0$. So we get

$$ \lim_{R \to \infty, \epsilon \to 0} \int_\Gamma f(z)\ dz = (1 - e^{2 \pi i (n+1)/m}) \int_0^\infty \frac{x^n\ dx}{1+x^m}$$

The meromorphic function $f(z)$ has one singularity inside $\Gamma$, a pole at $z = e^{\pi i/m}$ where the residue is $- e^{\pi i (n+1)/m}/m$. So the residue theorem gives you

$$ \int_0^\infty \frac{x^n\ dx}{1+x^m} = \frac{- 2 \pi i e^{\pi i (n+1)/m}}{ m (1 - e^{2 \pi i (n+1)/m})} = \frac{\pi}{m} \csc\left(\frac{\pi(n+1)}{m}\right)$$

Robert Israel
  • 448,999
  • 3
    "$\Gamma_3$ comes in along the ray at angle 2π/m. Since $e^{(2πi/m)m}=1, \int_{\Gamma_3}(z) ,dz=−e^{2πi(n+1)/m}∫_{\Gamma_1}f(z), dz$". How is it that the integral over $\Gamma_3$ becomes a multiple of the integral over $\Gamma_1$? – john.abraham Nov 21 '13 at 18:30
  • Hi @RobertIsrael, I like the method of using a wedge contour - it obviously minimizes computation of residues, and the conversion from complex number (lots of Euler formulas) to real number as a final answer to the original, real integral is a much easier process. However, say there are 3 roots, and a wedge going from 0 argument to pi/2 is enough to enclose 1 of the poles, why do we insist on using a wedge of angle 2pi/3 = 120 degrees? If we use a wedge of angle pi/2, does the positive imaginary axis pose a problem? Thanks Professor Israel, – User001 Oct 23 '15 at 05:52
  • 4
    The wedge contour works if the integrals over the radial parts are related by symmetry: multiplying $z$ by $e^{i\theta}$, where $\theta$ is the angle of the wedge, multiplies the function by a constant. The choice of angle is dictated by the symmetry of the function. – Robert Israel Oct 23 '15 at 06:13
  • why you used the small arc at the origin? Our function doesnt have a singularity at zero, it is not easier just use two straight lines and the big arc? – Masacroso Sep 20 '18 at 10:37
  • 2
    @Masacroso Habit, I guess. But it might be useful if $n$ is not an integer. – Robert Israel Sep 20 '18 at 16:24
  • If I choose a $n=0$ and $m=1.5$ then $\Gamma_3$ will lie directly on a pole. How would we deal with this? Moreover, for all $1<m<2$ we would be crossing a branch cut. How is this legal :( – Max0815 Aug 16 '22 at 01:10
36

Contour Integration Approach

Assuming only that $m>0$ and $-1<n<m-1$, let $z=x^m$ with the exaggerated contour $\gamma$:

keyhole contour

$\gamma$ is actually tight above and below the positive real axis and around the origin and circles back at an arbitrarily large distance from the origin.

The part just above the positive real axis captures the integral. The part just below the positive real axis gets $-e^{2\pi i\frac{n-m+1}{m}}$ times the integral. The residue at $z=-1$ of $\frac{z^{\frac{n-m+1}{m}}}{1+z}$ is $e^{\pi i\frac{n-m+1}{m}}$. Putting this all together yields

$$ \frac1m\int_\gamma\frac{z^{\frac{n-m+1}{m}}}{1+z}\,\mathrm{d}z=\left(1-e^{2\pi i\frac{n-m+1}{m}}\right)\int_0^\infty\frac{x^n}{1+x^m}\,\mathrm{d}x $$ $$ \frac{2\pi i}{m}e^{\pi i\frac{n-m+1}{m}}=\left(1-e^{2\pi i\frac{n-m+1}{m}}\right)\int_0^\infty\frac{x^n}{1+x^m}\,\mathrm{d}x $$ $$ \frac{\pi}{m}=\sin\left(\pi\frac{n+1}{m}\right)\int_0^\infty\frac{x^n}{1+x^m}\,\mathrm{d}x $$ $$ \frac{\pi}{m}\csc\left(\pi\frac{n+1}{m}\right)=\int_0^\infty\frac{x^n}{1+x^m}\,\mathrm{d}x\tag1 $$


Relation to $\bf{\Gamma(\alpha)\Gamma(1-\alpha)}$

Setting $m=1$, $n=\alpha-1$, and using the substitution $s=tu$ yields Euler's Reflection Formula: $$ \begin{align} \Gamma(\alpha)\Gamma(1-\alpha) &=\int_0^\infty s^{\alpha-1}e^{-s}\,\mathrm{d}s\int_0^\infty t^{-\alpha}e^{-t}\,\mathrm{d}t\\ &=\int_0^\infty\int_0^\infty u^{\alpha-1}e^{-(tu+t)}\,\mathrm{d}u\,\mathrm{d}t\\ &=\int_0^\infty\frac{u^{\alpha-1}}{1+u}\,\mathrm{d}u\\[6pt] &=\pi\csc(\pi\alpha)\tag2 \end{align} $$


Another Proof of $\bf{(1)}$ $$ \begin{align} \int_0^\infty\frac{x^{\alpha-1}}{1+x}\,\mathrm{d}x &=\int_0^1\frac{x^{-\alpha}+x^{\alpha-1}}{1+x}\,\mathrm{d}x\\ &=\sum_{k=0}^\infty(-1)^k\int_0^1\left(x^{k-\alpha}+x^{k+\alpha-1}\right)\mathrm{d}x\\ &=\sum_{k=0}^\infty(-1)^k\left(\frac1{k-\alpha+1}+\frac1{k+\alpha}\right)\\ &=\sum_{k\in\mathbb{Z}}\frac{(-1)^k}{k+\alpha}\\[9pt] &=\pi\csc(\pi\alpha)\tag3 \end{align} $$ where the last step of $(3)$ is proven in $(3)$ of this answer. Then apply $(3)$ to get $$ \begin{align} \int_0^\infty\frac{x^n}{1+x^m}\,\mathrm{d}x &=\frac1m\int_0^\infty\frac{x^{\frac{n-m+1}m}}{1+x}\,\mathrm{d}x\\ &=\frac\pi{m}\csc\left(\pi\frac{n+1}m\right)\tag4 \end{align} $$

robjohn
  • 345,667
  • Hi @robjohn, when faced with p.v. integrals, I always try to use this keyhole contour as my first method -- but if I remember correctly, this method has only ever worked nicely and correctly...once. Every other time, I've had to use a wedge contour, similar to what Robert Israel shows above. I wanted to ask you: are there times when the keyhole contour is the clearly better choice of contour - perhaps even necessary? If not, then using the wedge contour minimizes computation of residues and seems like the safest method to go with, especially on an exam, with the clock ticking. – User001 Oct 23 '15 at 05:43
  • What do you think? Thanks @robjohn, – User001 Oct 23 '15 at 05:43
  • Hi @robjohn, just a heads up, I've posted this comment as a new question on MSE. Thanks, – User001 Oct 23 '15 at 06:13
  • @robjohn , Hi ! May you detailled, ( I BEG YOU ) how can you compute the integral beneath the real positive axis ? I understand that the above part gives the original integral, but it appears to me that this is the same for the part under... where is my mistake and how to do you fix this ? THANK YOU. – Marine Galantin Jan 01 '19 at 17:14
  • 1
    @MarineGalantin: There is a branch cut along the real axis. $z^{\frac{n-m+1}m}$ limits to the usual real value from above the real axis, and it limits to $e^{2\pi i,\frac{n-m+1}m}$ times the usual real value from below the real axis. – robjohn Jan 01 '19 at 23:41
  • Why is there only 1 pole inside the contour? – Max0815 Aug 16 '22 at 06:02
  • @Max0815: there is a pole at $z=-1$. There is no other pole inside the contour. – robjohn Aug 16 '22 at 07:46
  • How so though? We have poles at every $e^{2\pi i/m}$. – Max0815 Aug 16 '22 at 07:59
  • The function is $\frac{z^{\frac{n-m+1}{m}}}{1+z}$. There is a branch cut from $0$ to $\infty$ and a pole at $-1$. – robjohn Aug 16 '22 at 08:01
  • Oh that makes sense, thanks – Max0815 Aug 16 '22 at 08:04
14

Use Ramanujan's Master theorem, with $u=x^b\Rightarrow dx=\frac{du}{bu^{1-\frac{1}{b}}}$:

$$\int_0^{\infty}\frac{x^{a-1}}{1+x^b}dx=\frac{1}{b}\int_0^{\infty}\frac{u^{\frac{a-1}{b}-1+\frac{1}{b}}}{1+u}du$$ $$=\frac{1}{b}\int_0^{\infty}u^{\frac{a}{b}-1}\sum_{k \ge 0}\frac{(-u)^k}{\Gamma(k+1)}\Gamma(k+1)du$$ $$=\frac{1}{b}\Gamma\left(\frac{a}{b}\right)\Gamma\left(1-\frac{a}{b}\right)=\frac{\pi}{b \sin \left(\frac{a\pi}{b}\right)}.$$

Meow
  • 6,353
14

$\newcommand{\+}{^{\dagger}} \newcommand{\angles}[1]{\left\langle #1 \right\rangle} \newcommand{\braces}[1]{\left\lbrace #1 \right\rbrace} \newcommand{\bracks}[1]{\left\lbrack #1 \right\rbrack} \newcommand{\ceil}[1]{\,\left\lceil #1 \right\rceil\,} \newcommand{\dd}{{\rm d}} \newcommand{\ds}[1]{\displaystyle{#1}} \newcommand{\equalby}[1]{{#1 \atop {= \atop \vphantom{\huge A}}}} \newcommand{\expo}[1]{\,{\rm e}^{#1}\,} \newcommand{\fermi}{\,{\rm f}} \newcommand{\floor}[1]{\,\left\lfloor #1 \right\rfloor\,} \newcommand{\half}{{1 \over 2}} \newcommand{\ic}{{\rm i}} \newcommand{\iff}{\Longleftrightarrow} \newcommand{\imp}{\Longrightarrow} \newcommand{\isdiv}{\,\left.\right\vert\,} \newcommand{\ket}[1]{\left\vert #1\right\rangle} \newcommand{\ol}[1]{\overline{#1}} \newcommand{\pars}[1]{\left( #1 \right)} \newcommand{\partiald}[3][]{\frac{\partial^{#1} #2}{\partial #3^{#1}}} \newcommand{\pp}{{\cal P}} \newcommand{\root}[2][]{\,\sqrt[#1]{\,#2\,}\,} \newcommand{\sech}{\,{\rm sech}} \newcommand{\sgn}{\,{\rm sgn}} \newcommand{\totald}[3][]{\frac{{\rm d}^{#1} #2}{{\rm d} #3^{#1}}} \newcommand{\ul}[1]{\underline{#1}} \newcommand{\verts}[1]{\left\vert\, #1 \,\right\vert}$ With $\tt @Pedro Tamaroff$ identity $\ds{\sum_{k = - \infty }^\infty{\pars{-1}^{k} \over a + k}}$ we'll have:

\begin{align} &\sum_{k = - \infty }^\infty{\pars{-1}^{k} \over a + k} = {1 \over a} + \sum_{k = 1}^{\infty}\bracks{ {\pars{-1}^{-k} \over a - k} + {\pars{-1}^{k} \over a + k}} = {1 \over a} + \sum_{k = 1}^{\infty}\pars{-1}^{k + 1} {2a \over \pars{k - a}\pars{k + a}} \\[5mm]&= {1 \over a} + \sum_{k = 0}^{\infty}{2a \over \pars{2k + 1 - a}\pars{2k + 1 + a}} - \sum_{k = 0}^{\infty}{2a \over \pars{2k + 2 - a}\pars{2k + 2 + a}} \\[5mm] = &\ {1 \over a} + {a \over 2} \sum_{k = 0}^{\infty}{1 \over \pars{k + \pars{1 - a}/2}\pars{k + \pars{1 + a}/2}} \\[2mm] &\ - {a \over 2}\sum_{k = 0}^{\infty} {1 \over \pars{k + \pars{1 - a/2}}\pars{k + \pars{1 + a/2}}} \\[5mm] = &\ {1 \over a} + {a \over 2}\,{\Psi\pars{\bracks{1 - a}/2} - \Psi\pars{\bracks{1 + a}/2} \over \pars{1 - a}/2 - \pars{1 + a}/2} - {a \over 2}\,{\Psi\pars{1 - a/2} - \Psi\pars{1 + a/2} \over \pars{1 - a/2} - \pars{1 + a/2}} \\[5mm] = &\ {1 \over a} + \half\bracks{ -\Psi\pars{1 - a \over 2} + \Psi\pars{1 + a \over 2} + \Psi\pars{1 -{a \over 2}} -\Psi\pars{1 + {a \over 2}}} \end{align}

With the identities: $$ \Psi\pars{1 - z} = \Psi\pars{z} + \pi\cot\pars{\pi z}\,,\quad \Psi\pars{\half + z} = \Psi\pars{\half - z} + \pi\tan\pars{\pi z}\,,\quad \Psi\pars{1 + z} = \Psi\pars{z} + {1 \over z} $$ we'll get

\begin{align} &\color{#0000ff}{\large \sum_{k = - \infty }^\infty{\pars{-1}^{k} \over a + k}} = {1 \over a} + \half\bracks{ \pi\tan\pars{\pi a \over 2} + \Psi\pars{a \over 2} + \pi\cot\pars{\pi a \over 2} - \Psi\pars{a \over 2} - {2 \over a}} \\[3mm]&= {\pi \over 2}\,{\sec^{2}\pars{\pi a/2} \over \tan\pars{\pi a/2}} = {\pi \over \sin\pars{\pi a}} = \color{#0000ff}{\large\pi\,\csc\pars{\pi a}} \end{align}
Felix Marin
  • 89,464
13

This is in Gradshteyn/Ryzhik, formula 3.241.2: $$\int_0^\infty {x^{\mu-1}\over 1+x^\nu}\ dx={\pi\over\nu}\csc\bigg({\mu\pi\over\nu}\bigg)={1\over\nu}B\bigg({\mu\over\nu},{\nu-\mu\over\nu}\bigg)$$ assuming, of course, $Re(\nu)>Re(\mu)>0$, and where $B(x,y)$ denotes the Beta function $B(x,y)={\Gamma(x)\Gamma(y)\over\Gamma(x+y)}$.


To see the Beta function part of the equality, use the integral representation $$B(x,y)=\int_0^\infty {t^{x-1}\over (1+t)^{x+y}}\ dt$$ Then ${1\over\nu}B\big({\mu\over\nu},{\nu-\mu\over\nu}\big)$ is $${1\over\nu}\int_0^\infty {t^{{\mu\over\nu}-1}\over 1+t}\ dt$$ Send $t$ to $t^\nu$, and you are done!

B R
  • 968
7

The answers on both of these previous threads

Simpler way to compute a definite integral without resorting to partial fractions?

$\int_{0}^{\infty}\frac{dx}{1+x^n}$

provide a solution to exactly this problem.

Also this problem arise again in this newer thread:

Is there an elementary method for evaluating $\int_0^\infty \frac{dx}{x^s (x+1)}$?

The meta thread discussing the similarities between these 4 questions can be found here: http://meta.math.stackexchange.com/questions/3746/abstract-duplicates-what-to-do-in-this-scenario

Eric Naslund
  • 72,099
  • Note that the previous qustions don't ask for the general solution but rather a particular case, and it is the answerers that decide to provide a solution to the general integral I present here. Also, I wanted another approach in the solution, see here. – Pedro Feb 29 '12 at 23:48
6

$$t=\frac1{1+x^m}\qquad\iff\qquad x=\left(\frac1t-1\right)^\frac1m=\left(\frac{1-t}t\right)^\frac1m=t^{^{-\frac1m}}\cdot(1-t)^{^\frac1m}\qquad\iff$$

Can you already see where this is going ? ;-)

$$\iff\frac{dx}{dt}=\frac{d}{dt}\left[\left(\frac1t-1\right)^\frac1m\ \right]=\frac1m\left(\frac{1-t}t\right)^{\frac1m-1}\left(-\frac1{t^2}\right)=-\frac{t^{^{-\frac1m-1}}\cdot(1-t)^{^{\frac1m-1}}}m\iff$$

$$I=\int_0^\infty\frac{x^n}{1+x^m}dx=\frac1m\int_0^1t^{^{-\frac{n+1}m}}\cdot(1-t)^{^{\frac{n+1}m-1}}dt=\frac{B\left(1-\frac{n+1}m,\frac{n+1}m\right)}m$$

Now all we have to do is use the fact that $B(m,n)=\displaystyle\frac{\Gamma(m)\Gamma(n)}{\Gamma(m+n)}$ , along with Euler's reflection formula, $\displaystyle\Gamma(z)\cdot\Gamma(1-z) = \frac{\pi}{\sin{(\pi z)}}$ , and we're done ! :-)

Lucian
  • 48,334
  • 2
  • 83
  • 154
3

Letting $u =x^m$ entails $dx =\frac{1}{m}t^{1/m-1}$ and hence,

$$ \int\limits_0^\infty {\frac{{{x^n}}}{{1 + {x^m}}}dx } =\int\limits_0^\infty {\frac{t^{\frac{n+1}{m}-1}}{(1 + t)^{1}}dt } =\frac{1}{m}B\left(\frac{n+1}{m}, 1-\frac{n+1}{m}\right)\\=\color{red}{\frac{1}{m}\Gamma\left(\frac{n+1}{m}\right)\Gamma\left( 1-\frac{n+1}{m}\right)=\frac{\pi}{m\sin\left(\frac{\pi(n+1)}{m}\right)}}$$

where we made use of the beta and Gamma function and the Schwartz duplication formula

Guy Fsone
  • 23,903
2

Integrate over a closed contour that is in the shape of a wedge in the 1st quadrant, having an angle of $\pi/5$. That is, consider

$$\oint_C dz \frac{z^2}{z^{10}+1}$$

where $C$ is that wedge contour. As mentioned above, the contour splits into 3 pieces:

$$\oint_C dz \frac{z^2}{z^{10}+1} = \int_0^R dx \frac{x^2}{x^{10}+1} + i R \int_0^{\pi/5}d\phi \, e^{i \phi} \frac{R^2 e^{i 2 \phi}}{R^{10} e^{i 10 \phi}+1}+ e^{i 3 \pi/5} \int_R^0 dx \frac{x^2}{x^{10}+1} $$

As $R \to \infty$, the second integral vanishes as $1/R^7$ (Note that the denominator never vanishes). The rest is then just equal to $i 2 \pi$ times the residue at the pole $z=e^{i \pi/10}$. Thus, combining the integrals, we have

$$\left (1-e^{i 3 \pi/5}\right) \int_0^{\infty} dx \frac{x^2}{x^{10}+1} = i 2 \pi \frac{e^{i 2 \pi/10}}{10 e^{i 9 \pi/10}} $$

Doing the algebra, I get

$$\int_0^{\infty} dx \frac{x^2}{x^{10}+1} = \frac{\pi}{10 \cos{(\pi/5)}}$$

Ron Gordon
  • 138,521
0

Letting $\displaystyle \frac{1}{t}=x^{m}+1$, then $ \displaystyle d x=\frac{1}{m} \left(\frac{1}{t}-1\right)^{\frac{1}{m}-1}\left(-\frac{1}{t^{2}}\right) d t. $

Consequently $$ \begin{aligned} \int_{0}^{\infty} \frac{x^{r}}{x^{m}+1} d x &=-\int_{1}^{0} \frac{\left(\frac{1}{t}-1\right)^{\frac{r}{m}}}{\frac{1}{t}} \frac{1}{m t^{2}}\left(\frac{1}{t}-1\right)^{\frac{1}{m}-1} d t \\ &=\frac{1}{m} \int_{0}^{1} \frac{(1-t)^{\frac{r}{m}}(1-t)^{\frac{1}{m}-1}}{t^{\frac{r+1}{m}}} d t \\ &=\frac{1}{m} \int_{0}^{1} t^{-\frac{r+1}{m}}(1-t)^{\frac{r+1}{m}-1} d t \\ &=\frac{1}{m} B\left(1-\frac{r+1}{m}, \frac{r+1}{m}\right) \end{aligned} $$ By the property of Beta function,

$$ B(x, y)=\frac{\Gamma(x) \Gamma(y)}{\Gamma(x+y)}, $$ where $\operatorname{Re}(x)>0$ and $\operatorname{Re}(y)>0,$

we have $$ \int_{0}^{\infty} \frac{x^{r}}{x^{m}+1} d x=\frac{\Gamma\left(1-\frac{r+1}{m}\right) \Gamma\left(\frac{r+1}{m}\right)}{m\Gamma(1)} $$

Using the Euler’s Reflection Theorem,

$$ \Gamma(1-z) \Gamma(z)=\pi \csc (\pi z), $$

where $z\notin Z$,

we can now conclude that $$\boxed{\int_{0}^{\infty} \frac{x^{r}}{x^{m}+1} d x=\frac{\pi}{m} \csc \frac{(r+1) \pi}{m}}$$

Lai
  • 20,421