Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

I would like to calculate $$\int_{0}^{\infty} \frac{ x^2 \log(x) }{1 + x^4}$$ by means of the Residue Theorem. This is what I tried so far: We can define a path $\alpha$ that consists of half a half-circle part ($\alpha_r$) and a path connecting the first and last point of that half circle (with radius $r$) so that we have $$ \int_{-r}^{r} f(x) dx + \int_{\alpha_r} f(z) dz = \int_{\alpha} f(z) dz = 2 \pi i \sum_{v = 1}^{k} \text{Res}(f;a_v) $$ where $a_v$ are zeros of the function $\frac{x^2 \log(x) }{1+x^4}$.

If we know $$\lim_{r \to \infty} \int_{\alpha_r} f(z) dz = 0 \tag{*} $$ then we know that $$\lim_{r \to \infty} \int_{-r}^{r} f(x) dx = \int_{-\infty}^{\infty} f(x) dx = 2 \pi i \sum_{v=1}^{k} \text{Res}(f;a_v) $$ and it becomes 'easy'.

Q: How do we know (*) is true?

share|cite|improve this question
up vote 11 down vote accepted

It's a bit more tricky that what you describe, but the general idea is correct. Instead of integrating from $0$ to $\infty$, one can integrate from $-\infty$ to $+\infty$ slightly above the real axis. Because of the logarithm, the integral from $-\infty$ to $0$ will give a possibly non-zero imaginary part, but the real part will be an even function of $x$. So we can write: \begin{align} \int_0^{\infty}\frac{x^2\ln x}{1+x^4}dx&=\frac12\mathrm{Re}\,\int_{-\infty+i0}^{\infty+i0} \frac{x^2\ln x}{1+x^4}dx=\\&=\pi\cdot \mathrm{Re}\left[ i\left(\mathrm{res}_{x=e^{i\pi/4}}\frac{x^2\ln x}{1+x^4}+\mathrm{res}_{x=e^{3i\pi/4}}\frac{x^2\ln x}{1+x^4}\right)\right]=\\ &=\pi\cdot \mathrm{Re}\left[ i\left(\frac{\pi e^{i\pi/4}}{16}- \frac{3\pi e^{3i\pi/4}}{16}\right)\right]=\\ &=\pi\cdot\mathrm{Re}\frac{(1+2i)\pi}{8\sqrt{2}}=\frac{\pi^2}{8\sqrt{2}}. \end{align}

Now as far as I understand the question was about how can one justify the vanishing of the integral over the half-circle $C$ which in its turn justifies the application of residue theorem. Parameterizing that circle as $x=Re^{i\varphi}$, $\varphi\in(0,\pi)$, we see that \begin{align} \int_C \frac{x^2\ln x}{1+x^4}dx=\int_0^{\pi}\frac{iR^3e^{3i\varphi}\left(i\varphi+\ln R\right)}{1+R^4e^{4i\varphi}}d\varphi=O\left(\frac{\ln R}{R}\right), \end{align} which indeed vanishes as $R\rightarrow \infty$.

share|cite|improve this answer
intresting way how to handle the branch cut. Very nice! – tired May 27 '15 at 15:36

Recalling the Mellin transform of a function $f$

$$ F(s)=\int_{0}^{\infty}x^{s-1}f(x)dx \implies F'(s) = \int_{0}^{\infty} x^{s-1}\ln(x)f(x) dx.$$

So, taking $f(x)=\frac{1}{x^4+1}$ and finding its Mellin transform

$$ F(s)=\frac{1}{4}\,{\frac {\pi }{\sin \left( \frac{\pi \,s}{4} \right) }} \implies F'(s)=\frac{1}{16}\,{\frac {{\pi }^{2}\cos \left( \pi \,s/4 \right) }{ \left( \cos \left( \pi \,s/4 \right) \right) ^{2}-1}}$$

Taking the limit as $s\to 3$ yields the desired result

$$ \lim_{s\to 3}F'(s)=\frac{\pi^2}{8\sqrt{2}}. $$

share|cite|improve this answer
The best way (+1) – user 1618033 Jul 10 '13 at 14:29

$$\underbrace{\dfrac{x^2 \log(x)}{1+x^4} dx \to \dfrac1{x^2} \dfrac{\log(1/x)}{1+1/x^4} \dfrac{-dx}{x^2}}_{x \to 1/x}$$ Hence, \begin{align} I = \int_0^{\infty} \dfrac{x^2 \log(x)}{1+x^4} dx & = \int_0^1 \dfrac{x^2 \log(x)}{1+x^4} dx + \int_1^{\infty} \dfrac{x^2\log(x)}{1+x^4}dx\\ & = \int_0^1 \dfrac{x^2 \log(x)}{1+x^4} dx - \int_0^1 \dfrac{\log(x)}{1+x^4}dx\\ & = \int_0^1 \dfrac{x^2-1}{1+x^4} \log(x) dx \end{align} \begin{align} \int_0^1 \dfrac{x^2 \log(x)}{1+x^4} dx & = \int_0^1 \sum_{k=0}^{\infty}(-1)^k x^{4k+2} \log(x)dx\\ & = \sum_{k=0}^{\infty}(-1)^k \int_{0}^{\infty} x^{4k+2} \log(x)dx\\ & = \sum_{k=0}^{\infty} \dfrac{(-1)^{k+1}}{(4k+3)^2} \end{align} \begin{align} \int_0^1 \dfrac{\log(x)}{1+x^4} dx & = \int_0^1 \sum_{k=0}^{\infty}(-1)^k x^{4k} \log(x)dx\\ & = \sum_{k=0}^{\infty}(-1)^k \int_{0}^{\infty} x^{4k} \log(x)dx\\ & = \sum_{k=0}^{\infty} \dfrac{(-1)^{k+1}}{(4k+1)^2} \end{align} $$I = \sum_{k=0}^{\infty} (-1)^k \left(\dfrac1{(4k+1)^2} - \dfrac1{(4k+3)^2}\right) = \dfrac{\pi^2}{8\sqrt2}$$ which can be obtained from the identity that the PolyLogarithmic function satisfies $$\text{Li}_n(e^{2 \pi ix}) + (-1)^n \text{Li}_n(e^{-2 \pi ix}) = - \dfrac{(2\pi i)^n}{n!}B_n(x)$$ You can see the posts here and here for more details.

share|cite|improve this answer
"By means of the residue theorem"? – mrf Jun 16 '13 at 19:48
@mrf The question just says "How to calculate $ \int_{0}^{\infty} \frac{ x^2 \log(x) }{1 + x^4} $?" and in the post, the OP makes a statement, he would like to calculate it "By means of the residue theorem" (which is not a question). So my post does answer his question. – user17762 Jun 16 '13 at 20:13

Another way to handle these kinds of integrals is by setting $$ f(z) = \frac{z^2 \log^2 z}{1+z^4} $$ where $\log$ is the "natural branch" of the complex logarithm and integrate $f$ along a keyhole contour, see for example this. The idea with this approach is to take advantage of the branch cut of $\log$.

share|cite|improve this answer

We can use Integration under Differentiation to solve/generalize this problem instead of residues or special functions. It is much easier. Let $$ J_n(a)=\int_0^\infty \frac{x^a}{1+x^n}dx. $$ Clearly $$ J_n'(a)=\int_0^\infty \frac{x^2\ln x}{1+x^n}dx. $$ Note $J_4'(0)=I$. First we calculate \begin{eqnarray} J_n(a)&=&\int_0^1\frac{x^a+x^{n-a-2}}{1+x^n}dx\\ &=&\int_0^1\sum_{k=0}^\infty(-1)^n(x^a+x^{n-a-2})x^{kn}dx\\ &=&\sum_{k=0}^\infty(-1)^k\left(\frac{1}{nk+a+1}+\frac{1}{nk+n-a-1}\right)\\ &=&\sum_{k=0}^\infty(-1)^k\frac{1}{nk+a+1}\\ &=&\sum_{k=-\infty}^\infty(-1)^k\frac{1}{nk+a+1}\\ &=&\frac{\pi}{n\sin(\frac{a+1}{n}\pi)}. \end{eqnarray} So $$ J_n'(a)=-\frac{\pi^2\cos(\frac{(a+1)\pi}{n})}{n^2\sin^2(\frac{(a+1)\pi}{n})} $$ and hence $$ I=J_4'(2)=\frac{\pi^2}{8\sqrt2}. $$ Here we used the following result $$ \sum_{k=-\infty}^\infty(-1)^k\frac{1}{ak+b}=\frac{\pi}{a\sin(\frac{b}{a}\pi)}.$$

share|cite|improve this answer

Use the substitution $x^4=t \Rightarrow dx=dt/(4t^{3/4})$ to obtain the following definite integral: $$\frac{1}{16}\int_0^{\infty} \frac{t^{-1/4}\ln t}{1+t}\,dt$$ Consider the result I proved here: What is $\int_0^\infty\!\!\int_0^\infty\!\!\int_0^\infty\!\frac{(xyz)^{-1/7}(yz)^{-1/7}z^{-1/7}}{(x+1)(y+1)(z+1)}\mathrm{d}x\mathrm{d}y\mathrm{d}z$ which states: $$\int_0^{\infty} \frac{t^{-a}}{1+t}\,dt=\frac{\pi}{\sin (a\pi)}$$ Differentiate the above with respect to $a$ to get: $$\int_0^{\infty} \frac{t^{-a}\ln t}{1+t}dt=\pi^2\cot(a\pi)\csc(a\pi)$$ Substitute $a=1/4$ and divide both the sides by 16 to get: $$\frac{1}{16}\int_0^{\infty} \frac{t^{-1/4}\ln t}{1+t}dt=\frac{\pi^2}{8\sqrt{2}}$$

share|cite|improve this answer

$\newcommand{\+}{^{\dagger}} \newcommand{\angles}[1]{\left\langle\, #1 \,\right\rangle} \newcommand{\braces}[1]{\left\lbrace\, #1 \,\right\rbrace} \newcommand{\bracks}[1]{\left\lbrack\, #1 \,\right\rbrack} \newcommand{\ceil}[1]{\,\left\lceil\, #1 \,\right\rceil\,} \newcommand{\dd}{{\rm d}} \newcommand{\down}{\downarrow} \newcommand{\ds}[1]{\displaystyle{#1}} \newcommand{\expo}[1]{\,{\rm e}^{#1}\,} \newcommand{\fermi}{\,{\rm f}} \newcommand{\floor}[1]{\,\left\lfloor #1 \right\rfloor\,} \newcommand{\half}{{1 \over 2}} \newcommand{\ic}{{\rm i}} \newcommand{\iff}{\Longleftrightarrow} \newcommand{\imp}{\Longrightarrow} \newcommand{\isdiv}{\,\left.\right\vert\,} \newcommand{\ket}[1]{\left\vert #1\right\rangle} \newcommand{\ol}[1]{\overline{#1}} \newcommand{\pars}[1]{\left(\, #1 \,\right)} \newcommand{\partiald}[3][]{\frac{\partial^{#1} #2}{\partial #3^{#1}}} \newcommand{\pp}{{\cal P}} \newcommand{\root}[2][]{\,\sqrt[#1]{\vphantom{\large A}\,#2\,}\,} \newcommand{\sech}{\,{\rm sech}} \newcommand{\sgn}{\,{\rm sgn}} \newcommand{\totald}[3][]{\frac{{\rm d}^{#1} #2}{{\rm d} #3^{#1}}} \newcommand{\ul}[1]{\underline{#1}} \newcommand{\verts}[1]{\left\vert\, #1 \,\right\vert} \newcommand{\wt}[1]{\widetilde{#1}}$ $\ds{\int_{0}^{\infty}{x^{2}\ln\pars{x} \over 1 + x^{4}}\,\dd x:\ {\large ?}}$

The following method 'avoids' the $\ds{\large\it\underline{\mbox{four poles}}}$ of the original integral.

Instead, we just have to consider a $\ds{\large\it\underline{\mbox{one pole-integral}}}$:

Note that \begin{align}&\color{#c00000}{\int_{0}^{\infty}{\ln\pars{x} \over 1 + x^{4}} \,\dd x} =\int_{0}^{\infty}{x^{1/2}\ln\pars{x^{1/4}} \over 1 + x}\,{1 \over 4}\,x^{-3/4}\,\dd x ={1 \over 16}\int_{0}^{\infty}{x^{-1/4}\ln\pars{x} \over 1 + x}\,\dd x \\[3mm]&={1 \over 16}\lim_{\mu\ \to\ -1/4}\partiald{}{\mu}\color{#00f}{% \int_{0}^{\infty}{x^{\mu} \over 1 + x}\,\dd x}\tag{1} \end{align}

\begin{align} &\color{#00f}{\int_{0}^{\infty}{x^{\mu} \over 1 + x}\,\dd x} =2\pi\ic\expo{\pi\mu\ic} - \int_{\infty}^{0} {x^{\mu}\expo{2\pi\mu\ic} \over 1 + x}\,\dd x \\[3mm]&\imp\quad \color{#00f}{\int_{0}^{\infty}{x^{\mu} \over 1 + x}\,\dd x} =2\pi\ic\,{\expo{\pi\mu\ic} \over 1 - \expo{2\pi\mu\ic}} =-\pi\,{2\ic \over \expo{\pi\mu\ic} - \expo{-\pi\mu\ic}} =\color{#00f}{-\,{\pi \over \sin\pars{\pi\mu}}} \end{align}

Replacing in $\pars{1}$: \begin{align}&\color{#66f}{\large\int_{0}^{\infty}{\ln\pars{x} \over 1 + x^{4}} \,\dd x} ={1 \over 16}\bracks{\pi^{2}\cot\pars{\pi\mu}\csc\pars{\pi\mu}}_{\mu\ =\ -1/4} =\color{#66f}{\large {\root{2} \over 16}\,\pi^{2}} \approx -0.8724 \end{align}

share|cite|improve this answer

Your Answer


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

Not the answer you're looking for? Browse other questions tagged or ask your own question.