7

How can we find the integral: $$\int_{ - \infty }^{ + \infty } {\frac{1} {1 + {x^4}}} \;{\mathrm{d}}x$$ I tried to find and got it to be $\cfrac{\pi}{\sqrt2}$. Am I correct? Please help me with an appropriate method. I tried to use sum of residue.

  • 1
    The residue theorem is fine. What were your problems with that? Another option is to transform it into something you recognise as an integral representation of the beta function. – Daniel Fischer Oct 22 '14 at 13:38
  • I believe the standard approach is to consider a contour integral over the semicircle with radius $R$ and then show that $f(z)$ vanishes along the arc. One then applies the residue theorem by considering the two poles that lie within the contour. – Gahawar Oct 22 '14 at 13:46
  • Or find a good change of variable, without using advanced and frankly unappropriate complex analysis tools – mvggz Oct 22 '14 at 13:48
  • @mvggz I would rather beat my horse with an imaginary flail than a wooden stick. – Gahawar Oct 22 '14 at 13:51
  • Wait for it, I'm giving you something interesting :) – mvggz Oct 22 '14 at 14:00
  • @mvggz I see that one could have the opinion that the residue theorem is advanced in some way, but what makes you think complex analysis methods weren't appropriate? – Daniel Fischer Oct 22 '14 at 14:02
  • I'm saying this because you can deal with this in a few ligns, with only very basic knowledge regarding change of variable. I've always been taught to use theorems only when there is no other choice. In a certain way, you should be able to prove everything you use, so the simpler your tools are the stronger your proof is. Do you see what I mean? The residue theorem is almost too powerful for that type of question – mvggz Oct 22 '14 at 14:07
  • @mvggz Once you have the residue theorem firmly under your belt, there is no reason not to use it when it makes something easy. The same holds for other theorems of course. If you're a complex analyst, evaluating that integral with the residue theorem is very likely the easiest way, or at least one of the two or three easiest ways, for you. If you barely know what a holomorphic function is, then it is of course not appropriate for you. (The "you" is of course generic here, not addressing you in person.) There are of course a myriad (well, at least half a dozen or so) good ways to evaluate it. – Daniel Fischer Oct 22 '14 at 14:25
  • Yes I agree, but it's a general way of doing things. If I can answer a question with tools as simple as possible, and yet elegantly :) , I'll do it. For instance complex analysis is not studied before the end of third year in France (I'm french..), while what I used to anwser is the mere basis of integration. It's the way I was taught, if I answered a question with sophisticated and very powerful theorems when I could have done it using very little material, I was heavily penalized. My teacher used to tell me that you don't smash a ladybug with an atomic bomb, this is of course a metaphor – mvggz Oct 22 '14 at 14:33

7 Answers7

4

$\newcommand{\angles}[1]{\left\langle\, #1 \,\right\rangle} \newcommand{\braces}[1]{\left\lbrace\, #1 \,\right\rbrace} \newcommand{\bracks}[1]{\left\lbrack\, #1 \,\right\rbrack} \newcommand{\ceil}[1]{\,\left\lceil\, #1 \,\right\rceil\,} \newcommand{\dd}{{\rm d}} \newcommand{\ds}[1]{\displaystyle{#1}} \newcommand{\dsc}[1]{\displaystyle{\color{#c00000}{#1}}} \newcommand{\expo}[1]{\,{\rm e}^{#1}\,} \newcommand{\fermi}{\,{\rm f}} \newcommand{\floor}[1]{\,\left\lfloor #1 \right\rfloor\,} \newcommand{\half}{{1 \over 2}} \newcommand{\ic}{{\rm i}} \newcommand{\iff}{\Longleftrightarrow} \newcommand{\imp}{\Longrightarrow} \newcommand{\Li}[1]{\,{\rm Li}_{#1}} \newcommand{\pars}[1]{\left(\, #1 \,\right)} \newcommand{\partiald}[3][]{\frac{\partial^{#1} #2}{\partial #3^{#1}}} \newcommand{\pp}{{\cal P}} \newcommand{\root}[2][]{\,\sqrt[#1]{\vphantom{\large A}\,#2\,}\,} \newcommand{\sech}{\,{\rm sech}} \newcommand{\sgn}{\,{\rm sgn}} \newcommand{\totald}[3][]{\frac{{\rm d}^{#1} #2}{{\rm d} #3^{#1}}} \newcommand{\ul}[1]{\underline{#1}} \newcommand{\verts}[1]{\left\vert\, #1 \,\right\vert}$ \begin{align}&\color{#66f}{\large\int_{-\infty}^{\infty}{\dd x \over x^{4} + 1}} =2\int_{0}^{\infty}{\dd x \over x^{2} + 1/x^{2}}\,{1 \over x^{2}}\,\dd x\ =\ \underbrace{\overbrace{% 2\int_{0}^{\infty}{1 \over \pars{x - 1/x}^{2} + 2}\,{1 \over x^{2}}\,\dd x} ^{\ds{\color{#c00000}{x\ \mapsto\ {1 \over x}}}}}_{\ds{\color{#00f}{{\cal I}_{0}}}} \\[5mm]&=2\int_{\infty}^{0}{1 \over \pars{1/x - x}^{2} + 2}\,x^{2}\, \pars{-\,{\dd x \over x^{2}}} =\underbrace{2\int_{0}^{\infty}{1 \over \pars{x - 1/x}^{2} + 2}\,\dd x} _{\ds{\color{#00f}{{\cal I}_{1}}}} \end{align}

As $\ds{\color{#00f}{{\cal I}_{0}}\ \mbox{and}\ \color{#00f}{{\cal I}_{1}}}$ are equal to the original integral we'll have:

\begin{align}&\color{#66f}{\large\int_{-\infty}^{\infty}{\dd x \over x^{4} + 1}} ={\color{#00f}{{\cal I}_{0}} + \color{#00f}{{\cal I}_{1}} \over 2} =\int_{0}^{\infty}{1 \over \pars{x - 1/x}^{2} + 2}\,\pars{1 + {1 \over x^{2}}} \,\dd x \\[5mm]&=\overbrace{\int_{x\ =\ 0}^{x\ \to\ \infty}{1 \over \pars{x - 1/x}^{2} + 2} \,\dd\pars{x - {1 \over x}}} ^{\ds{\color{#c00000}{t \equiv x - {1 \over x}}}} =\int_{-\infty}^{\infty}{\dd t \over t^{2} + 2} =\root{2}\ \overbrace{\int_{0}^{\infty}{\dd t \over t^{2} + 1}} ^{\ds{\color{#c00000}{\pi \over 2}}} =\color{#66f}{\large{\root{2} \over 2}\,\pi} \end{align}

Felix Marin
  • 89,464
4

Since the integrand is an even function, we can rewrite it as \begin{equation} \int_{-\infty}^\infty\dfrac{1}{1+x^4}\ dx=2\int_{0}^\infty\dfrac{1}{1+x^4}\ dx \end{equation} In general, we have (click the formula below for the proof) \begin{equation} \int_0^\infty\dfrac{1}{1+x^n}\ dx=\frac{\pi}{n\sin\left(\frac{\pi}{n}\right)},\qquad\mbox{for}\qquad n>0 \end{equation} For the residue approach, you might refer to this link.

3

Consider contour, which is semicircle (origin at $0$ and radius $R$), more precisely:

$$\gamma=\{Re^{it}:t \in [0,\pi] \} \cup [-R,R]$$

According residue theorem:

$$\oint_{\gamma}\frac{1}{x^4+1}dz=2\pi i (\text{Res}_{b}\frac{1}{x^4+1}+\text{Res}_{a}\frac{1}{x^4+1})$$

Where $a=e^{\pi i/4}$ and $b=e^{2\pi i /4}$

But on the other hand:

$$\oint_{\gamma}\frac{1}{x^4+1}dz=\int_{-R}^{R}\frac{1}{x^4+1}dz+\int_{\gamma'(R)}\frac{1}{x^4+1}dz$$

where $\gamma'(R)=\{Re^{it}:t \in [0,\pi] \}$.

Lets try to calculate $\int_{\gamma'(R)}\frac{1}{x^4+1}dz$, for example using simple parametrisation:

$$\int_{\gamma'(R)}\frac{1}{x^4+1}dz=i\int_{0}^{\pi}Re^{it}\frac{1}{1+R^4e^{4it}}dt$$

It's clear that:

$$\left|Re^{it}\frac{1}{1+R^4e^{4it}}\right| \to 0$$

when $R \to \infty$ for $t \in [0,\pi]$, so when $R \to \infty$:

$$0=\lim_{R \to \infty}\int_{\gamma'(R)}\frac{1}{x^4+1}dz=\lim_{R \to \infty}\oint_{\gamma}\frac{1}{x^4+1}dz-\int_{-R}^{R}\frac{1}{x^4+1}dz= \\ = \oint_{\gamma}\frac{1}{x^4+1}dz-\int_{-\infty}^{\infty}\frac{1}{x^4+1}dz$$

agha
  • 10,038
2

First, use the change of variable $x=\frac{1}{u} => dx = -\frac{1}{u^2}du $

$I = \int_{ - \infty }^{ + \infty } {\frac{1} {1 + {x^4}}} \;{\mathrm{d}}x = 2*\int_{ 0 }^{ + \infty } {\frac{1} {1 + {x^4}}} \;{\mathrm{d}}x = 2*\int_{ 0 }^{ + \infty } {\frac{u^2} {1 + {u^4}}} \;{\mathrm{d}}u $

=> $2*I = 2*\int_{ 0 }^{ + \infty } {\frac{1+x^2} {1 + {x^4}}} \;{\mathrm{d}}x => I= \int_{ 0 }^{ + \infty } {\frac{1+x^2} {1 + {x^4}}} \;{\mathrm{d}}x $

Now: Let: $t= x-\frac{1}{x}$ => $dt= \frac{1+x^2}{x^2}dx $ x->0, t->$-\infty$ ; x->$+\infty$ , t->$+\infty$

=> $ I = \int_{ - \infty }^{ + \infty } {\frac{1+x^2} {1 + {x^4}}*\frac{x^2}{1+x^2}} \;{\mathrm{d}}t = \int_{ - \infty }^{ + \infty } {\frac{1} {x^{-2} + {x^2}}} \;{\mathrm{d}}t$

$ x^2 + x^{-2} = (x-\frac{1}{x})^2 +2 = t^2 +2 $

=> $ I = \int_{ - \infty }^{ + \infty } {\frac{1} {2 + {t^2}}} \;{\mathrm{d}}t $

Let : t = $\sqrt{2}*v$

$I = \frac{1}{2}*\int_{ - \infty }^{ + \infty } {\frac{1} {1 + {(\frac{t}{\sqrt{2}})^2}}} \;{\mathrm{d}}t = \frac{1}{\sqrt{2}}*\int_{ - \infty }^{ + \infty } {\frac{1} {1 + {v^2}}} \;{\mathrm{d}}v = \frac{\pi}{\sqrt{2}}$

mvggz
  • 1,965
  • 1
    This proof can be adapted straightforwardly to any function of the form $\frac{1}{x^4 + a x^2 + b}$ or $\frac{x^2}{x^4 + a x^2 + b}$. – user111187 Oct 22 '14 at 19:28
1

One may be interested in the following solution: \begin{eqnarray} \int_{-\infty}^\infty\frac{1}{1+x^4}dx&=&2\int_{0}^\infty\frac{1}{1+x^4}dx\\ &=&-2\Im\int_{0}^\infty\frac{1}{x^2+i}dx\\ &=&-2\Im\frac{1}{\sqrt{i}}\arctan\frac{x}{\sqrt{i}}\bigg|_0^\infty\\ &=&-2\Im\frac{1}{\sqrt{i}}\frac{\pi}{2}\\ &=&-\pi\Im\frac{1}{\frac{\sqrt{2}}{2}+i\frac{\sqrt{2}}{2}}\\ &=&\frac{\pi}{\sqrt{2}}. \end{eqnarray}

xpaul
  • 44,000
1

use that $1+x^4=- \left( x\sqrt {2}-{x}^{2}-1 \right) \left( x\sqrt {2}+{x}^{2}+1 \right) $ and make a partial fraction decomposition

1

If I did this problem, I would first prove that $\int_{-\infty}^{\infty}\frac{1}{1+x^4}dx$ converges, then we have

$$\int_{-\infty}^{\infty}\frac{1}{1+x^4}dx=\lim_{R\rightarrow\infty}\int_{-R}^{R}\frac{1}{1+x^4}dx=\lim_{R\rightarrow\infty}\left[\int_{C(R)^+}\frac{1}{1+z^4}dz-\int_{L(R)^+}\frac{1}{1+z^4}dz\right]$$

Where $L(R)$ is the upper part of the circle with radius $R$ and $C(R)$ is the closed contour made by $L(R)$ and the segment $[-R,R]$. Then I would evaluate $\lim_{R\rightarrow\infty}\int_{C(R)^+}\frac{1}{1+z^4}dz$ by residual. The part $\lim_{R\rightarrow\infty}\int_{L(R)^+}\frac{1}{1+z^4}dz$ may be prove to be $0$ if we notice that $\left|\frac{1}{1+z^4}\right|\le\frac{2}{R^4}$ when $R$ approaches $\infty$.

anonymous67
  • 3,458