2

Chances are, I am missing something trivial here, but I ran into a problem when trying to calculate the following improper integral:

$$ I=\int_{-\infty}^{+\infty}\mathrm{d}x\frac{\exp(-x^2)}{x^2+1} . $$

My approach was to treat it as a complex integral, and use the Cauchy formula:

$$ I'=\int_{\Gamma}\mathrm{d}z\frac{\exp(-z^2)}{z^2+1}=2\pi i \left(\frac{\exp(-z^2)}{z+i}\right)_{z=i}=\pi e , $$

where $\Gamma$ is a closed curve (with the usual infinite-radius semicircle on the upper half-plane, denoted as $\gamma$). We can partition $I'$ as

\begin{align*} I'=& I+\int_{\gamma}\mathrm{d}z\frac{\exp(-z^2)}{z^2+1}= \\=& I+\underbrace{\lim_{R\rightarrow\infty}\int_0^{2\pi}\mathrm{d}\phi \frac{iR\exp(i\phi)\exp(-R^2\exp(i2\phi))}{R^2\exp(i2\phi)+1}}_{=0}\\ =&I . \end{align*}

Based on this, $I=\pi e$. However, the actual answer should be

$$ I=\pi e (1-\mathrm{erf}(1)) . $$

It seems as if I missed a $\pi e\mathrm{erf}(1)$ term on the right hand side, but I cannot see how...

I know the integral is evaluated in this SE post. But I would be grateful if someone could correct my Cauchy formula-based reasoning, if possible.

0 Answers0