1

Imagine you are given the following integral: $$ I = \int_{-\infty}^\infty\frac{\sin(x)}{x(x^2 + 1)} $$ If we transform the problem into a contour integral we have the following:

$$ \int_C \frac{e^{iz}}{z(z^2 + 1)}dz = 2 \pi i \sum_k \text{Res}(f, z_k) $$ if we perform the integration over the contour $C$, given by: $$ \gamma = [-R, -\varepsilon] \cup C_{\varepsilon} \cup [\varepsilon, R] \cup C_\Gamma $$ where $[z_1, z_2]$ denotes the segment that joins the points $z_1$ and $z_2$, and: $$ C_\varepsilon = \varepsilon e^{it}, \; \; t \in [\pi, 0] $$ $$ C_\Gamma = R e^{it}, \; \; t \in [0, \pi] $$ Further analyzation gives the following: $$ \int_C \frac{e^{iz}}{z(z^2 + 1)}dz = \lim_{(R, \varepsilon) \to (\infty, 0)}\int_{-R}^{-\varepsilon} + \int_{C_\varepsilon} + \int^{R}_{\varepsilon} = 2 \pi i \sum_k \text{Res}(f, z_k) $$ Hence, we finally have: $$ \int_{-\infty}^{\infty} \frac{e^{iz}}{z(z^2 + 1)}dz + \pi i \text{Res}(f, z = 0) = 2\pi i\text{Res}(f, z = i) $$ $$ \int_{-\infty}^{\infty} \frac{e^{iz}}{z(z^2 + 1)}dz = 2\pi i\text{Res}(f, z = i) - \pi i \text{Res}(f, z = 0) $$ where: $$ \text{Res}(f, z = i) = \frac{-1}{2e} $$ $$ \text{Res}(f, z = 0) = 1 $$

$$ \int_{-\infty}^{\infty} \frac{e^{iz}}{z(z^2 + 1)}dz = i \left ( -\pi - \frac{\pi}{e} \right) $$ $$ I = \text{Im} \left [ i \left ( -\pi - \frac{\pi}{e} \right) \right ] = -\pi - \frac{\pi}{e} $$ So the result of the integral is: $$ I = -\pi - \frac{\pi}{e} $$ which is wrong. I realised the term coming from the residue of $z = 0$ is wrong, and changing it sign would fix the result. I've calculate the residues with the formula and by calculating the $a_{-1}$ term of the Laurent Series the function. What mistake did I make? Is my line of reasoning wrong?

J. W. Tanner
  • 60,406
  • You are traversing $C_\varepsilon$ in the clockwise direction... – giobrach Feb 02 '24 at 17:17
  • @giobrach Ahhh..., so that would flip the sign of the integral (and the residue), and give me the correct value, no? – Álvaro Rodrigo Feb 02 '24 at 17:18
  • $0$ is outside of your contour. You took a semicircular path of radius $\epsilon$ to avoid it. – user317176 Feb 02 '24 at 17:26
  • Your question is very popular here. See https://math.stackexchange.com/questions/3790802/mistake-in-calculating-int-infty-infty-frac-sinxx1x2?noredirect=1, https://math.stackexchange.com/questions/2996303/contour-integration-evaluating-int-infty-infty-frac-sinxxx21?noredirect=1, https://math.stackexchange.com/questions/722891/contour-integration-part?noredirect=1, https://math.stackexchange.com/questions/2091924/calculate-int-infty-infty-frac-sin-tt1t2-mathrmdt?noredirect=1, https://math.stackexchange.com/questions/3911489/how-to-prove-int-0-infty-frac-sin-xxx21dx-frac-pi21-frac?noredirect=1. – Gonçalo Feb 02 '24 at 19:20

0 Answers0