1

I started off with this contour(I apologize for the software i used but I am new to this all).

i used this contour https://dochub.com/m/shared-document/madhavasthana/1XEpyxzwN5JN1vZVQZGd38/null-1-png?dt=wSzeX2E6NohcehhbdSE6

there is a branch cut from $i \to i{\infty}$, and another one from $-i \to -i{\infty}$.

the contour encloses only two poles, namely $z=e^{\pi i/4}$ and $z=e^{3\pi i/4}$.

by residue theorem, $$\oint_C \frac{\log(z^{2}+1)}{z^{4}+1}dz = 2 \pi i\sum_{z_i=e^{i\pi/4},e^{3i\pi/4}}\operatorname{Res}\left[\frac{\log(z^{2}+1)}{z^{4}+1},z_i\right]$$

upon applying L'Hospital rule to evaluate the residues,

$$\oint_C = 2\pi i(\lim_{z \to e^\frac{i\pi}{4}}(z-e^\frac {i\pi}{4})\frac{\log(z^{2}+1)}{z^{4}+1}+\lim_{z \to e^\frac{3\pi i}{4}}(z-e^\frac {3\pi i}{4})\frac{\log(z^{2}+1)}{z^{4}+1})=2\pi i\left(\frac{\log(1+i)}{4e^{\frac{3i\pi}{4}}} + \frac{\log(1-i)}{4e^{\frac{i\pi}{4}}}\right)=\frac{\pi}{2}[e^\frac{-i\pi}{4}(\frac{ln2}{2}+ i\frac{\pi}{4})+e^\frac{i\pi}{4}(\frac{ln2}{2}-i\frac{\pi}{4})]=\frac{\pi}{2}[\frac{{e^\frac{i\pi}{4}}+e^\frac{-i\pi}{4}}{2}ln2+(\frac{1}{\sqrt2}+\frac{i}{\sqrt2})(\frac{\pi}{4}-\frac{i\pi}{4})]=\frac{\pi}{2}[\frac{1}{\sqrt2}ln2+\frac{\pi}{4\sqrt2}(1+i)(1-i)]=\frac{\pi}{2\sqrt2}[ln2+\frac{\pi}{2}]$$

I am guessing I made a mistake in the arguments when opened them using $\log(z)=\log|z|+ i\arg z$, taking the arguments as $\frac{\pi}{4}$ for $z=1+i$ and $\frac{-\pi}{4}$ for $z=1-i$, but that yielded the wrong answer. apart from the LHS having the residues, everything in the RHS with the twice the desired integral and line integrals along the imaginary axis and around $z=i$ were right as follows

If $$\displaystyle \int_0^{ \infty } \frac{\log(x^{2}+1)}{x^{4}+1}dx=I$$

then in the limit as $\lim_{a \to \infty}$ and $\lim_{b \to 0}$

$$\frac{\pi}{2\sqrt2}[ln2+\frac{\pi}{2}]= 2I - 2\pi i \int_i^{i\infty}\frac{1}{z^{4}+1} dz$$ (the round trip around $z=i$, cancels out the $\log$ terms and leaves a difference of $2\pi i$).

I would appreciate some help on how to evaluate the desired integral with this contour.

  • Hi, welcome to Math SE. Hint: if you replace $\ln(x^2+1)$ with $2\Re\ln(x+i)$ before making the integration variable complex, you won't need a keyhole. – J.G. Feb 14 '24 at 08:31
  • I didn't think of that. This is my first time evaluating an integral with two branch cuts, so I didn't think of that, I will try it. But How do I evaluate the residue with the logarithms containing log(1+i) and log(1-i)? – Madhav Asthana Feb 14 '24 at 08:54
  • I'm not sure what you mean. If you make the integration range $\Bbb R$ using the fact the original integrand is even, then use my suggestion, the enclosed poles are at $e^{i\pi/4},,e^{3i\pi/4}$. Then, for example, the residue at the first is$$\lim_{z\to e^{i\pi/4}}\frac{z-e^{i\pi/4}}{z^4+1}\ln(z+i).$$The log factor comes outside, then you can use L'Hôpital's rule. – J.G. Feb 14 '24 at 09:02
  • your suggestion is better. I agree. Writing ln(x^2+1) and $\Re$ln(x+i) does indeed short the branch cut at i and removes the keyhole. But I am curious so as to how to simplify the residues obtained when you use a keyhole and open the log terms using log(z)=log|z|+i(argz), I know that |1$\pm$i|=2, but upon setting the arguments as $\pm \frac {\pi}{4}$, the answer obtained was wrong – Madhav Asthana Feb 14 '24 at 09:07
  • If you edit your question to show what happened when you tried the calculation, we can probably point to a mistake. – J.G. Feb 14 '24 at 09:12
  • i edited the answer to fit my calculations – Madhav Asthana Feb 14 '24 at 10:44
  • I found a video here and I have a somewhat sloppy and similar answer. I parameterized each subset of your contour into their real and imaginary parts and coded them here using geometry and $\operatorname{Arg}\left(1+z^2\right) \in (-\pi,\pi]$. If you click on that link, you can see why J.G.'s hint is way easier. I can write an answer using that keyhole contour if you want. – Accelerator Feb 15 '24 at 04:54
  • 1
    @Accelerator I did take J.G.'s approach and you are right. I resolved the problem quickly. but i was still curious as to how to evaluate the residues within my original contour. I would really appreciate it if you could take the time to write an answer using this contour https://www.desmos.com/calculator/tk3bcapobo – Madhav Asthana Feb 15 '24 at 06:49
  • I'll write an answer. I need around a week to do so since I'm busy. In the meantime, please feel free to read the answer I posted here, specifically the part on how I calculated the residue. Your issue might be how you split up the logarithm or your choice of arguments, but I'm not sure since I haven't read everything carefully yet. As for the circular arcs, you can use the ML-inequality and use the Desmos link to find the arc lengths and show their integrals go to zero. – Accelerator Feb 15 '24 at 19:38
  • @Accelerator I appreciate it – Madhav Asthana Feb 17 '24 at 17:52
  • @MadhavAsthana I've written a solution that uses your keyhole contour. All works nicely. Be careful about how the branches are defined and evaluating the ensuing complex arguments of $\log(z^2+1)$. Finally, the integral $\int_1^\infty \frac1{x^4+1},dx$ can be calculated directly since $\frac1{x^4+1}$ has a closed-form anti-derivative. Please let me know how I can improve my answer. I really want to give you the best answer I can. – Mark Viola Feb 19 '24 at 23:21

1 Answers1

1

There are several ways to proceed. Here, we will evaluate the integral of interest $I$, as given by

$$I=\int_0^\infty \frac{\log(x^2+1)}{x^4+1}\,dx$$

using contour integration. Hence, we will evaluate the conotu integral $J$ as given by

$$J=\oint_C \frac{\log(z^2+1)}{z^4+1}\,dz$$

where $C$ is the keyhole contour as defined by the OP. Using the residue theorem, we find that

$$\begin{align} \int_{-\infty}^\infty \frac{\log(x^2+1)}{x^4+1}\,dx&=\int_1^\infty \frac{[\log(x^2-1)+i\pi]-[\log(x^2-1)-i\pi]}{1+x^4}\,i\,dx\\\\ &+2\pi i \text{Res}\left(\frac{\log(z^2+1)}{z^4+1}, z=e^{i\pi/4}\right)\\\\ 2\int_0^\infty \frac{\log(x^2+1)}{x^4+1}\,dx&=-2\pi \int_1^\infty \frac1{x^4+1}\,dx+2\pi i \left(\frac{\log(\sqrt{2})+i\pi/4}{4e^{i3\pi/4}}+\frac{\log(\sqrt2)-i\pi/4}{4e^{i9\pi/4}}\right)\\\\ \int_0^\infty \frac{\log(x^2+1)}{x^4+1}\,dx&=- \underbrace{\int_1^\infty \frac{\pi}{x^4+1}\,dx}_{=\pi^2\sqrt{2}/8 -\pi\sqrt{2}\text{coth}^{-1}(\sqrt{2})/4}+\frac\pi 4\left(\sqrt2 \log(\sqrt2)+\pi \sqrt 2/4\right)\\\\ &=\frac{\pi \sqrt 2}{4}\left(\log(2+\sqrt2)-\pi/4\right) \end{align}$$

And we are done!

Mark Viola
  • 179,405
  • Thank you, good sir. I saw my mistake was not in the residues but rather the integral around the keyhole that leaves $$\int_{i}^{i\infty} \frac {-2\pi idz}{z^{4}+1}$$I made a fault in the substitution z=ix and missed an additional i, not accounting for the +ve sign that gets carried on to the LHS where the residues are – Madhav Asthana Feb 20 '24 at 16:07
  • You're welcome. My pleasure. – Mark Viola Feb 20 '24 at 16:36