Can anyone please give me a detailed solution of $\int_{-\infty}^{\infty}\frac{1}{1+3x+5x^2} dx$. I have tried looking at the solutions of other problems like these but I am unable to work this out. A simple and detailed solution that works for a lot of problems like these would be very helpful.
-
Which similar problems did you look at, and how were they done? – Michael Hardy Apr 10 '18 at 03:41
-
@MichaelHardy saw this https://math.stackexchange.com/questions/135917/calculating-a-real-integral-using-complex-integration – Kousik Krishnan Apr 10 '18 at 03:48
-
Partial fractions or residue calculus may help. – pshmath0 Apr 10 '18 at 09:05
1 Answers
Let $\gamma$ be a contour with two parts. $\gamma_1$ is the portion of real line between $[-R,R]$ And $\gamma_2$ is the semi-circular path of radius $R$ through the portion of the complex plain with positive real part.
$$\int_{-R}^R f(x) \, dx = \oint_\gamma f(z) \, dz - \int_{\gamma_2} f(z) \, dz $$
To evaluate $\oint_\gamma f(z) \, dz$ we just need to evaluate the residuals with positive imaginary part.
$$(5x^2 + 3x + 1) = 5\left(x + \frac {3 + i\sqrt {11}}{10}\right) \left(x + \frac {3 - i\sqrt {11}}{10}\right)$$
We have one residual in the region at $x = \frac {-3 + i\sqrt {11}}{10}$
Which will evaluate to $2\pi \frac {1}{\sqrt {11}}$
And what remains is to show that $\int_{\gamma_2} f(z) dz = 0$
$$z = Re^{it}\\ dz = Rie^{it}\\ \int_{\gamma_2} f(z) \, dz = \int_0^\pi \frac {Rie^{it}}{5R^2 e^{2it} + 3R e^{it} + 1} < \pi \frac {R}{5R^2}$$
Taking the limit a R goes to infinity, we see that this indeed equals 0.

- 57,877
-
@MichaelHardy Thanks, get so focused on the calculus that a person forgets their algebra. – Doug M Apr 10 '18 at 04:11