Let $\mu$ be a finite measure supported by $\Gamma $ (a smoth finite curve) and absolutely continuous with respect to the length measure on $\Gamma$ such that $\Gamma \cap (\Gamma+x)$ is a finite set for any non-zero $x\in \mathbb{R^2}.$

$\int_{R^2}\varphi(x)d\mu(x)=0$ ,for every $2\pi$-periodic function $\varphi$.

that means that the locally finite measure defined by the sum $\sum_{n\in \mathbb Z^2} d\mu(x-2\pi n)$.

is in fact the zero measure ???

  • $\begingroup$ Please make clear what question you are asking. $\endgroup$ Jan 3 '16 at 17:50
  • $\begingroup$ please why $\int_{R^2}\varphi(x)d\mu(x)=0$ implies $\sum_{n\in{\mathbb{Z}}^2} d\mu(x-2\pi n)=0$?thank you . $\endgroup$ Jan 3 '16 at 18:55
  • $\begingroup$ When you integrate against the measure defined by the sum, you're integrating the periodization against $d\mu$. In symbols, $\langle \phi, \chi_{\mathbb{Z}^2}\ast d\mu\rangle = \langle \chi_{\mathbb{Z}^2}\ast\phi, d\mu\rangle = 0$. $\endgroup$ Jan 3 '16 at 20:08
  • $\begingroup$ Thank you sir ,but what's this '∗' ,$\chi_{{\mathbb{Z}}^2}$? ,i need some explanations thank you so much $\endgroup$ Jan 3 '16 at 20:19
  • $\begingroup$ For other readers of this question: from comments on the OP's other question it seems that these questions relate to sciencedirect.com/science/article/pii/S0007449710000953 $\endgroup$
    – Yemon Choi
    Jan 5 '16 at 16:32

I don't think the definition of $\mu$ is relevant to your problem, you only need that $\int\phi d\mu =0 $ if $\phi$ is $2\pi$-periodic.

The reason is that the measure $$ d\nu(x) = \sum_{n\in\mathbb{Z}^2}d\mu(x-2\pi n) $$ (which is the convolution of the indicator function of the lattice $\chi_{2\pi\mathbb{Z}^2}$ with $d\mu$) is defined to make integration of $\phi$ against $d\nu$ the same as integration of the periodization of the $\phi$ against $d\mu$: $$ \int \phi(x)d\nu(x) =\int\phi(x)\left(\sum_{n\in\mathbb{Z}^2}d\mu(x-2\pi n)\right) = \int\left(\sum_{n\in\mathbb{Z}^2}\phi(y+2\pi n)\right)d\mu(y) $$ The periodization $$ \sum_{n\in\mathbb{Z}^2}\phi(y+2\pi n) $$ is $2\pi$-periodic, so the second integral is always zero.

(Throughout, let's assume that $\phi$ has enough decay to make all the integrals converge. Also, we need to exchange the order of the integral and the sum to make the change of variables $x\to y+2\pi n$, then change the order back.)

  • $\begingroup$ Thank you for your help and your patience but ,what is $d\mu(x)$ ? $\endgroup$ Jan 20 '16 at 23:53
  • $\begingroup$ It's any measure such that $\int\phi d\mu = 0$ if $\phi$ is $2\pi$-periodic, so in particular, your measure. $\endgroup$ Jan 21 '16 at 18:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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