1
$\begingroup$

I got the improper integral: $$ \int_0^\infty \frac{x^2}{x^4+1} \, dx $$

On one hand one could of course just evaluate the integral and then apply limits. However, it's not always practical to find an antiderivative. I got a hint that one can trace back improper integrals to sequences and thus just calculate the limit of a sequence which is way easier, could someone explain to me how this works or link me to some theorems? Thanks.

$\endgroup$
  • $\begingroup$ This looks like a standard candidate for evaluation by contour integration. $\endgroup$ – Travis Willse May 30 '17 at 14:44
  • $\begingroup$ Are you actually trying to evaluate it or just trying to show that it has a finite value? $\endgroup$ – tilper May 30 '17 at 14:44
  • $\begingroup$ I just need to show it has a finite value, and i haven't had contour integration sorr :/ $\endgroup$ – johnka May 30 '17 at 14:50
2
$\begingroup$

The value of this integral is finite because $f(x):={ \frac { x^ 2 }{ x^ 4+1 } }$ is continuous in $\mathbb{R}$ and $f(x)\sim 1/x^2$ as $x$ goes to $\pm\infty$.

It can be evaluated by using Residue Theorem: $$\int _{ 0 }^{ \infty }f(x)dx= \frac{1}{2}\int _{ -\infty }^{ \infty }f(x)dx= \pi i(\mbox{Res}(f,e^{i\pi/4})+\mbox{Res}(f,e^{i3\pi/4}))\\= \pi i\left(\frac{1}{4e^{i\pi/4}}+\frac{1}{4ie^{i\pi/4}}\right)= \frac{\pi}{2\sqrt{2}}.$$

Alternatively, you can integrate $f$ by noting that $$x^4+1 =(x^2+1)^2 - (\sqrt2x)^2 = (x^2+\sqrt2x+1)(x^2-\sqrt2x+1)$$ Then, by using partial fraction decomposition, we get $$f(x)=\frac{1}{2\sqrt{2}}\left(\frac{x}{x^2-\sqrt2x+1} -\frac{x}{x^2+\sqrt2x+1}\right).$$

$\endgroup$
  • $\begingroup$ Sorry I'm not familiar with complexe Analysis yet $\endgroup$ – johnka May 30 '17 at 14:54
  • $\begingroup$ Then try the second approach $\endgroup$ – Robert Z May 30 '17 at 14:58
  • $\begingroup$ I have already tried the second approch before but its fairly long, and i think i formulated my question wrong i only need to show the existence :) $\endgroup$ – johnka May 30 '17 at 15:22
  • $\begingroup$ @johnka For the convergence my first two lines suffice. $\endgroup$ – Robert Z May 30 '17 at 15:29
1
$\begingroup$

Based on the overall nature of your post (and the comment you left that I saw as I finished this), it sounds like you're trying to show this integral actually has a finite value and you're not necessarily trying to find what that value is.

Since the integrand is $\dfrac{x^2}{x^4+1}$, then we know that as $x$ gets really super huge (i.e., as $x \to +\infty$), the integrand "behaves like" $\dfrac{x^2}{x^4} = \dfrac1{x^2}$. And we know that $\displaystyle \int_1^{+\infty} \dfrac1{x^2} \, dx$ has a finite value. Actually we can calculate it directly, and it's $1$. (This does not mean the original integral is also $1$. We can't conclude anything about the original integral yet!)

So, because our integrand $\dfrac{x^2}{x^4+1}$ "behaves like" an integrand that has a finite integral over $[1,+\infty)$, we expect that our integral of $\dfrac{x^2}{x^4+1}$ over $[0,+\infty)$ is also finite.

To prove it, we can just use the direct comparison test. On $[1, +\infty)$, we know that $x^4 \le x^4 + 1$. Divide both sides by $x^2(x^4+1)$ to conclude that $\dfrac{x^2}{x^4+1} \le \dfrac1{x^2}$. Since $\displaystyle \int_1^{+\infty} \dfrac1{x^2} \, dx$ is finite, then so is $\displaystyle \int_0^{+\infty} \dfrac{x^2}{x^4+1} \, dx$.

Note 1: Technically the above directly implies that $\displaystyle \int_1^{+\infty} \dfrac{x^2}{x^4+1} \, dx$ is finite, but $\displaystyle \int_0^1 \dfrac{x^2}{x^4+1} \, dx$ is clearly finite and so we conclude $\displaystyle \int_0^1 \dfrac{x^2}{x^4+1} \, dx + \displaystyle \int_1^{+\infty} \dfrac{x^2}{x^4+1} \, dx = \displaystyle \int_0^{+\infty} \dfrac{x^2}{x^4+1} \, dx$ is finite.

Note 2: For the comparison test for integrals to work, you also need to note that $\dfrac{x^2}{x^4+1} \ge 0$ on $[0,+\infty)$ (technically you just need on $[1,+\infty)$). This is obviously true, but to be thorough it really should be at least mentioned.

$\endgroup$
  • $\begingroup$ Thanks, good answer. Could you argue with f is continious on the Intervall (0,1) and therefor you can integrate it Thus watching only the intervall (1,INF) because we know that the statement is true in the other fraction? $\endgroup$ – johnka May 30 '17 at 15:20
  • $\begingroup$ @johnka, yes, since $\dfrac{x^2}{x^4+1}$ is continuous on $[0,1]$, it has a maximum value on $[0,1]$. Say this maximum value is $M$. Then $\displaystyle \int_0^1 \dfrac{x^2}{x^4+1} \, dx \le M(1-0) = M$, and so all we really need to worry about is whether or not $\displaystyle \int_1^{+\infty} \dfrac{x^2}{x^4+1} \, dx$ is finite, and it is by the integral comparison test. $\endgroup$ – tilper May 30 '17 at 15:31
1
$\begingroup$

$$I=\int_{0}^{1}\frac{x^2}{x^4+1}\,dx + \int_{1}^{+\infty}\frac{x^2}{x^4+1}\,dx = \int_{0}^{1}\left(\frac{x^2}{x^4+1}+\frac{1}{x^4+1}\right)\,dx \tag{1}$$ leads to: $$ I = \int_{0}^{1}\frac{1+x^2-x^4-x^6}{1-x^8}\,dx = \sum_{k\geq 0}\left(\frac{1}{8k+1}+\frac{1}{8k+3}-\frac{1}{8k+5}-\frac{1}{8k+7}\right)\tag{2}$$ but due to the digamma reflection formula we have: $$ \sum_{k\geq 0}\left(\frac{1}{8k+1}-\frac{1}{8k+7}\right)=\frac{\pi}{8}\cot\frac{\pi}{8}, \quad \sum_{k\geq 0}\left(\frac{1}{8k+3}-\frac{1}{8k+5}\right)=\frac{\pi}{8}\cot\frac{3\pi}{8}\tag{3}$$ hence it follows that: $$ I = \frac{\pi}{8}\left[\cot\frac{\pi}{8}+\cot\frac{3\pi}{8}\right]=\color{red}{\frac{\pi}{2\sqrt{2}}}.\tag{4}$$ The same can be achieved by partial fraction decomposition, since $x^4+1=\Phi_8(x)=(x^2-x\sqrt{2}+1)(x^2+x\sqrt{2}+1)$.

Still another way is to apply Glasser's master theorem: $$ \int_{0}^{+\infty}\frac{dx}{x^2+\frac{1}{x^2}} = \int_{0}^{+\infty}\frac{dx}{\left(x-\frac{1}{x}\right)^2+2}=\int_{0}^{+\infty}\frac{dx}{x^2+2}.\tag{5} $$

$\endgroup$
1
$\begingroup$

Note that $$I=\int_{0}^{\infty}\frac{x^{2}}{x^{4}+1}dx\stackrel{x^{4}=u}{=}\frac{1}{4}\int_{0}^{\infty}\frac{u^{-1/4}}{u+1}du$$ and now we can see that this is one of the form of the Beta function. So $$I=\frac{B\left(\frac{3}{4},\frac{1}{4}\right)}{4}=\color{red}{\frac{\pi}{2\sqrt{2}}}.$$

$\endgroup$
1
$\begingroup$

$\newcommand{\bbx}[1]{\,\bbox[15px,border:1px groove navy]{\displaystyle{#1}}\,} \newcommand{\braces}[1]{\left\lbrace\,{#1}\,\right\rbrace} \newcommand{\bracks}[1]{\left\lbrack\,{#1}\,\right\rbrack} \newcommand{\dd}{\mathrm{d}} \newcommand{\ds}[1]{\displaystyle{#1}} \newcommand{\expo}[1]{\,\mathrm{e}^{#1}\,} \newcommand{\ic}{\mathrm{i}} \newcommand{\mc}[1]{\mathcal{#1}} \newcommand{\mrm}[1]{\mathrm{#1}} \newcommand{\pars}[1]{\left(\,{#1}\,\right)} \newcommand{\partiald}[3][]{\frac{\partial^{#1} #2}{\partial #3^{#1}}} \newcommand{\root}[2][]{\,\sqrt[#1]{\,{#2}\,}\,} \newcommand{\totald}[3][]{\frac{\mathrm{d}^{#1} #2}{\mathrm{d} #3^{#1}}} \newcommand{\verts}[1]{\left\vert\,{#1}\,\right\vert}$

Lets $\ds{\mc{C}_{R}}$ a quarter circumference in the first quadrant with radius $\ds{R > 1}$ and vertex at the origin. Such contour 'encloses' a single pole $\ds{p \equiv \expo{\ic\pi/4}}$.

\begin{align} \Re\oint_{\mc{C}_{R}}{z^{2} \over z^{4} + 1}\,\dd z & = \Re\pars{2\pi\ic\,\lim_{z \to p}{\bracks{z - p}z^{2} \over z^{4} - 1}} = -2\pi\,\Im\pars{3p^{2} - 2p^{2} \over 4p^{3}} \\[5mm] & = -\,{1 \over 2}\,\pi\,\Im\pars{1 \over p} = -\,{1 \over 2}\,\pi\,\Im\pars{\expo{-\ic\pi/4}} = {\root{2} \over 4}\,\pi\label{1}\tag{1} \\[1cm] \Re\oint_{\mc{C}_{R}}{z^{2} \over z^{4} + 1}\,\dd z & = \int_{0}^{R}{x^{2} \over x^{4} + 1}\,\dd x \\[3mm] & +\ \underbrace{\Re\pars{\int_{0}^{\pi/2}{R^{2}\expo{2\ic\theta} \over R^{4}\expo{4\ic\theta} + 1}\,R\expo{\ic\theta}\ic\,\dd\theta}} _{\ds{\to\ 0\ \mrm{as}\ R\ \to\ \infty}}\ +\ \underbrace{% \Re\pars{\int_{R}^{0}{-y^{2} \over y^{4} + 1}\,\ic\,\dd y}}_{\ds{=\ 0}} \label{2}\tag{2} \end{align}

\eqref{1} and \eqref{2} lead, in the $\ds{R \to \infty}$ limit, to

$$ \bbx{\int_{0}^{\infty}{x^{2} \over x^{4} + 1}\,\dd x = {\root{2} \over 4}\,\pi} $$


ANOTHER METHOD: \begin{align} \int_{0}^{\infty}{x^{2} \over x^{4} + 1}\,\dd x & = \int_{0}^{\infty}{\dd x \over x^{2} + 1/x^{2}} = \int_{0}^{\infty}{\dd x \over \pars{x - 1/x}^{2} + 2} \\[5mm] & = {1 \over 2}\bracks{% \int_{0}^{\infty}{\dd x \over \pars{x - 1/x}^{2} + 2} + \int_{\infty}^{0}{1 \over \pars{1/x - x}^{2} + 2}\pars{-\,{\dd x \over x^{2}}}} \\[5mm] & = {1 \over 2} \int_{0}^{\infty}{1 + 1/x^{2} \over \pars{x - 1/x}^{2} + 2} \,\dd x \,\,\,\stackrel{t\ =\ x - 1/x}{=}\,\,\, {1 \over 2}\int_{-\infty}^{\infty}{\dd t \over t^{2} + 2} \\[5mm] & = {1 \over \root{2}}\int_{0}^{\infty}{\dd t \over t^{2} + 1} = {\root{2} \over 2}\,{\pi \over 2} = \bbx{{\root{2} \over 4}\,\pi} \end{align}

$\endgroup$

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.