1
$\begingroup$

Let $K(x):=|x|^{-\alpha}$ be a function on $\mathbb{R}^{n}\setminus\{0\}$ with $0<\alpha<n$. Suppose $\phi$ is a $C^{\infty}_{c}(\mathbb{R}^n)$ function such that

$$\text{(i)}\quad \text{supp}\phi \subset \left\{\frac{1}{2}<|x|<2\right\},$$ $$\text{(ii)}\quad\sum_{k\in \mathbb{Z}} \phi(2^{-k}x)=1.$$

Let $$g(\xi):=\mathcal{F}\left(\sum_{k=1}^{\infty}\phi(2^{-k}x)K \right)(\xi),$$ where $\mathcal{F}$ denotes Fourier transform.

Problem: show that $$g(\xi)=\sum_{k=1}^{\infty} \mathcal{F}\left(\phi(2^{-k}x)K \right)(\xi)$$ with convergence in the space of tempered distributions.

Given a Schwartz function $\psi$, we need to show that $$ \sum_{k=1}^{m}\langle \mathcal{F}\left(\phi(2^{-k} x)|x|^{-\alpha}\right),\psi\rangle \to\langle \mathcal{F}\left(\sum_{k=1}^{\infty}\phi(2^{-k} x)|x|^{-\alpha}\right), \psi\rangle.$$ Equivalently, we need to prove that

$$ \sum_{k=1}^{m}\langle \phi(2^{-k} x)|x|^{-\alpha},\widehat{\psi}\rangle=\langle \sum_{k=1}^{m}\phi(2^{-k} x)|x|^{-\alpha},\widehat{\psi}\rangle\\ \to \langle \sum_{k=1}^{\infty}\phi(2^{-k} x)|x|^{-\alpha}, \widehat{\psi}\rangle.$$

My argument is simple: $$\sum_{k=1}^{m}\phi(2^{-k} x)\to \sum_{k=1}^{\infty}\phi(2^{-k} x)$$ which is a bounded continuous function (Recall that $\phi$ is smooth and $\sum_{k=-\infty}^{\infty}\phi(2^{-k} x)=1$). Also, $|\cdot|^{-\alpha} \widehat{\psi} $ is an $L^1$ function. So the passing of the limit is now justified by the dominated convergence theorem.

Did I get the question correctly ? And if so is the solution correct ?

$\endgroup$
3
  • 2
    $\begingroup$ The details actually don't matter, this follows from the fact that the FT is continuous on $\mathcal S'$. $\endgroup$ Jan 28, 2023 at 23:30
  • 1
    $\begingroup$ Which can be easily proved: If $T_{n}\to T$ in $\mathcal{S}^{\prime}$ then $$\langle \mathcal{F}T_{n},\phi\rangle=\langle T_{n},\mathcal{F}\phi \rangle\to \langle T,\mathcal{F} \phi \rangle=\langle \mathcal{F}T,\phi \rangle$$ for every $\phi \in \mathcal{S}$. $\endgroup$
    – Medo
    Jan 28, 2023 at 23:47
  • $\begingroup$ Thank you, prof. Remling. $\endgroup$
    – Medo
    Jan 28, 2023 at 23:50

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.