0
$\begingroup$

I'm stuck on a question involving evaluating improper integrals using the residue theorem.
Here's what I'm trying to evaluate:
$\int_{-\infty}^{\infty} \frac{1}{(1+x^2)^3} dx$

To begin with, we can define a contour integral over the the semi-circle disc above the Real axis called $C$ which can be broken up in two parts involving the Real number line, and the curve of the semi-circular disc
$\int_{C} \frac{1}{(1+z^2)^3} dz = \int_{-R}^{R} \frac{1}{(1+x^2)^3} dx + \int_{Cr} \frac{1}{(1+z^2)^3} dz$

We can see that the left-hand side of the equation has poles at $i$ and $-i$ with multiplicity $3$ respectively. Using the Residue Theorem
$\int_{C} \frac{1}{(1+z^2)^3} dz = 2 \pi i (3 \frac{1}{((i)+i)^3})$

Because there are $3$ instances of the $i$ being the pole inside the region defined.
Also:
$\int_{Cr} \frac{1}{(1+z^2)^3} dz = 0$
Because parameterizing $z = Re^{it}$ and setting the limit as R goes to infinity will show the integrand to be zero.

That leaves us with the final answer of
$-\frac{6 \pi}{8}$

But this is not correct.
The correct answer is $\frac{3 \pi}{8}$ and I'm not quite sure how that is formed.

Any help would be appreciated.
Thanks.

$\endgroup$
1
$\begingroup$

The residue of $\dfrac1{(1+z^2)^3}$ at $i$ is $-\dfrac{3i}{16}$ and therefore your integral is equal to$$2\pi i\times\frac{-3i}{16}=\frac{3\pi}8$$indeed.

In order to compute that residue, you can use the fact that$$\dfrac1{(1+z^2)^3}=\frac{\frac1{(z+i)^3}}{(z-i)^3}$$and that therefore the residue that we are interested in is $\dfrac{\varphi''(i)}2$, where $\varphi(z)=\dfrac1{(z+i)^3}$. And, as I wrote, $\dfrac{\varphi''(i)}2=-\dfrac{3i}{16}$.

$\endgroup$
  • $\begingroup$ Why can't we find the residue by multiplying the original function by $(z-1)^3$, isn't that how you find a residue usually? $\endgroup$ – jd94 Dec 13 '18 at 17:39
  • $\begingroup$ Why $(z-1)^3$? The number $1$ has nothing to do with this problem. It should be $(z-i)^3$, and that is exactly what I did. $\endgroup$ – José Carlos Santos Dec 13 '18 at 17:51
  • $\begingroup$ Typo on my end. So the idea is, if we have a higher order pole, we differentiate it until the order of it is 1, and divide by how many times we differentiated factorial? $\endgroup$ – jd94 Dec 13 '18 at 17:56
  • $\begingroup$ I would not put it like that, but that's the right idea. $\endgroup$ – José Carlos Santos Dec 13 '18 at 18:08
0
$\begingroup$

By the Cauchy integral formula

if $a$ is inside the contour, and $f(z)$ is holomorphic inside the contour.

$\int_\gamma \frac {f(z)}{z-a} dz = 2\pi i f(a)$

and

$\int_\gamma \frac {f(z)}{(z-a)^n} dz = 2\pi i \frac {1}{(n-1)!} f^{(n-1)}(a)$

in this case let $a = i$

and $f(z) = \frac {1}{(z+i)^3}$

$\endgroup$

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.