Re: Extracurricular Integration Marathon
This integral is clearly a prime target for contour integration. (We could factorise into its real quadratic factors, use partial fractions, and integrate the simpler summands, but contour integration seems faster).
Note that since the quartic denominator is both even and real, the poles of the integrand occur at
where
is an arbitrarily chosen root, say the one in the first quadrant of the the complex plane.
Now if we take a semicircular contour (radius R) with diameter on the real axis centred at the origin, and semicircular arc in the upper half-plane, then the integral around this contour (with positive orientation) is just equal to I, because the integrand decays as 1/R^2, and so the contribution from the curved segment of length O(R) tends to zero.
On the other hand, for sufficiently large R this integral is just equal to
It remains to compute
, which is just the principal square root of
.
Writing
, we get
The resulting biquadratic yields
, where
is the golden ratio.
Hence