0

Here is the problem:

Contour Integral problem

With help from Jack D'Aurizio

We were able to prove that the contour integral of the big semi circle $=0$ as $R \to \infty$.

Now the problem is the small semi-circle.

How do I prove as $\delta \to 0$ that the contour integral of the small semi-circle $=0$?

I can try to eliminate the $\log$

We start with

$$\left | \int f \right| < \int |f| $$

$$|f| < \frac{\sqrt{\log^2(\delta) - \theta^2}}{\delta^2 - 1}$$

We take the limit as $\delta \to 0$, the RHS $\to -\infty$ which is no good.

But we see that, $x^2 > \log^2(x)$

So:

$$|f| < \frac{\sqrt{\log^2(\delta) - \theta^2}}{\delta^2 - 1} < \frac{\sqrt{\delta^2 - \pi^2}}{\delta^2 - 1} $$

But that is still no good.

What we can however do is prove that:

$$x^2 > \log^2(x) - \pi^2$$

It is already true that: for x > 0

$$x^2 > \log^2(x)$$ $$0 > -\pi^2$$

$$\therefore x^2 > \log^2(x) - \pi^2$$

The limit if $x^2$ is there will be 0.

But will the estimation lemma work?

Amad27
  • 10,465
  • Parametrise using $z=\epsilon e^{i\phi}$ $$\int^{\pi}_0\frac{\ln^2(\epsilon e^{i\phi})}{1+\epsilon^2e^{i2\phi}}i\epsilon^{i\phi}d\phi$$ and take the limit as $\epsilon\to0$. – M.N.C.E. Dec 24 '14 at 16:18
  • @M.N.C.E., how are we allowed to bring the limit INSIDE? – Amad27 Dec 24 '14 at 17:44

0 Answers0