Consider the Littlewood-Paley decomposition of a tempered distribution $f$: \begin{equation} f = \tilde{\theta}(D)f + \sum_{k \geq 1} \theta(2^{-k}D)f. \end{equation} Here $(\tilde{\theta}, \theta)$ are smooth compactly supported functions ($\theta$ having a support contained in an annulus far from zero), satisfying the following dyadic partition of unity relation: \begin{equation} \tilde{\theta}(\xi) + \sum_{k \geq 1} \theta(2^{-k}\xi) = 1, \quad \forall \xi \in \mathbb{R}. \end{equation} In addition, $\tilde{\theta}(D)$ (as well as $\theta(2^{-k}D)$) are Fourier multipliers (i.e., they are defined by $\mathcal{F}(\tilde{\theta}(D)f)(\xi) = \tilde{\theta}(\xi)f(\xi)$ for all $\xi \in \mathbb{R}$). For $0 < \alpha < 1$, consider $C^{0, \alpha}(\mathbb{R})$ the space of continous and bounded functions $f: \mathbb{R} \rightarrow \mathbb{R}$ satisfying the Holder condition, i.e. \begin{equation} \exists c > 0 : \forall (x, y) \in \mathbb{R} \times \mathbb{R}, \quad |f(x) - f(y)| \leq C |x-y|^{\alpha}. \end{equation} Assume that $f$ is continous and bounded and \begin{equation} \| f \|_{C^{0, \alpha}}^{LP} := \| \tilde{\theta}(D)f\|_{L^{\infty}} + \sup_{k \in \mathbb{N}} 2^{k\alpha} \|\theta(2^{-k}D)f \|_{L^{\infty}} < +\infty. \end{equation} I am looking to prove the following estimates. Fix $x, y \in \mathbb{R}$ and let $N \in \mathbb{N}$ be arbitrary. First,

\begin{equation} \sum_{k > N} 2 \|\theta(2^{-k}D)f\|_{L^{\infty}} \leq \frac{\| f \|_{C^{0, \alpha}}^{LP}}{2^{\alpha}-1} \frac{1}{2^{\alpha N}}. \end{equation}


\begin{equation} \Big|\tilde{\theta}(D)f(x) - \tilde{\theta}(D)f(y) + \sum_{k=1}^{N}[\theta(2^{-k}D)f(x) - \theta(2^{-k}D)f(y)] \Big| \leq C|x-y|\sum_{k=1}^{N} 2^{k(1-\alpha)}\| f \|_{C^{0, \alpha}}^{LP} \end{equation}

for some $C > 0$. Concering the first estimate, I "went in brutally" and I manage to obtain a bound like $\| f \|_{C^{0, \alpha}}^{LP} \sum_{k > N} 2^{1-k\alpha}$, so I don't know how to go further. For the second estimate, I sense an impending use of semi-classical Bernstein inequalities but I fail to see how to apply it to obtain such a bound. Any hint is welcomed.


I have succeeded (sort of, modulo a factor 2) in proving the first estimate. Recall that when $|q| < 1$, the remainder of the associated geometric series writes \begin{equation} \sum_{k=n}^{\infty} q^k = \frac{q^n}{1-q}. \end{equation} Thus, \begin{align} \sum_{k>N}^\infty 2 \| \theta(2^{-k}D)f\|_{L^{\infty}} &= 2\sum_{k = N+1}^\infty \frac{2^{k\alpha}}{2^{k\alpha}} \| \theta(2^{-k}D)f\|_{L^{\infty}} \\ &\leq 2 \sup_{k \in \mathbb{N}} 2^{k\alpha} \| \theta(2^{-k}D)f\|_{L^{\infty}} \sum_{k=N+1}^\infty \frac{1}{2^{k\alpha}} \\ &\leq 2\|f\|_{C^{0, \alpha}}^{LP} \frac{2^{-(N+1)\alpha}}{1-2^{-\alpha}} = \frac{2\|f\|_{C^{0, \alpha}}^{LP}}{2^{\alpha}-1}\frac{1}{2^{\alpha N}}. \end{align} I still don't know how to prove the second inequality.

Edit: I may have found a way to prove the second inequality (using a mean value inequality then applying some of the hypotheses) - I'll write it down as soon as I have some time.


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.