Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Hello, I'm trying to bound an integral. I have a function $A(\nu) = | 1 + \exp(-I \nu) |$ (with $I$ being the imaginary unit) and I want to show that the condition (Paley-Wiener criterion for causality) applies

$$\int_{-\infty}^{\infty} \frac{|\log(A(\omega))|}{1+\omega^2} \mathrm{d}\omega < \infty$$

(log is the natural logarithm) I used a transformation from $\omega$ to $\nu$: $\omega = \tan(\nu/2)$ and I converted the integral by substitution to

$$\int_{-\pi}^{\pi} | \log(A(\nu)) |\mathrm{d}\nu < \infty$$

But I don't know how to show that this condition applies for the given function $A(\nu)$. I tried to simplify the problem by using $A(\nu) = | 1 + \exp(-I \nu) | = \sqrt{(1+\exp(-I\nu))(1+\exp(I\nu))}$ and thus simplifying the integral to

$$\frac{1}{2} \int_{-\pi}^{\pi} | \log(1+\exp(-I\nu)) + \log(1+\exp(I\nu)) |\mathrm{d}\nu$$

But still I have trouble finding a bound. I also tried $A(\nu) = | 1 + \exp(-I \nu) | = \sqrt{2} \sqrt{\cos(\nu) + 1}$.

I though maybe the problem can be solved by providing an upper and lower bound function that converges. But because $A(\nu)$ has values in the range $[0,1]$ the logarithm assumes very large values (and there are actually points of singularity for $A(\nu)=0$).

Please help me solve this problem.

share|improve this question

2 Answers 2

up vote 2 down vote accepted

You want to show that $$\int_{-\pi}^\pi|\log|1+e^{-it}||dt$$ is finite. Now $$|1+e^{-it}|=|e^{it/2}+e^{-it/2}|=2\cos(t/2)$$ so your integral is $$\int_{-\pi}^\pi|\log|2\cos(t/2)||dt =2\int_0^\pi|\log|2\cos(t/2)||dt.$$ Replacing $t$ by $\pi-2$ in the last integral gives $$2\int_0^\pi|\log|2\sin(t/2)||dt.$$ The integrand is nicely continuous away from $0$. Near $0$, $\sin (t/2)=tf(t)$ where $f(t)\to1/2$ as $t\to0$. Then the integrand is $|\log t+g(t)|$ where $g$ is continuous at $0$ and now finiteness follows since $$\int_0^1|\log t|dt$$ is finite (integration by parts).

share|improve this answer
    
Thank you very much for your answer. (I think you got a small typo. Replacing $t$ by $\pi-t$ and not $\pi-2$) But I have troubles following the last step of your argument. Why can I ignore the $g(t)$ for the proof of finiteness? –  user10256 Oct 22 '10 at 16:02
    
oh I think I understand it now. $log t$ is not continuous at $0$. But $g(t)$ is. So I don't have to worry about $g(t)$. It has no singularities and therefore is finite. The question left is whether the singularity of $log t$ causes the integral to be infinite. But it is easy to show that $\int_0^1 | \log t | \mathrm{d}t$ is finite and therefore the initial integral. Thanks thanks thanks. –  user10256 Oct 23 '10 at 21:52

The key is to understand the behavior of $A(\nu)$ near the singularity $\nu=0$. Using Taylor expansion we know that for $\nu$ small $A(\nu) = 1+e^{-I \nu} \approx -I\nu$. Therefore, $\log|A(\nu)| \approx \log|-I \nu| = \log |\nu|$. Note that $\int_{-\pi}^\pi \log|\nu| d\nu = 2 \int_{0}^\pi \log \nu d\nu < \infty$. To make this precise you need to control the error terms in the Taylor approximation.

share|improve this answer
    
There is no singularity at $\nu=0$, but at $\nu=\pm\pi$. For $nu$ small, $A(\nu)\approx2$. –  Julián Aguirre Oct 22 '10 at 14:58
    
Oops, as Julian points out I made a mistake above. To handle the singularities at $\pi$ and $-\pi$ you can still use a Taylor approximation. Near $\pi$ we have $A(\nu) \approx -I(\nu-\pi)$. Therefore the integral should be finite if $\int_0^\pi \log|\nu-\pi|\, d\nu < \infty$. –  Jon Peterson Oct 25 '10 at 14:15

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.