Characterizations of Besov-type and Triebel-Lizorkin-type spaces via maximal functions and local means (Q607097)

Let \(S(\mathbb{R}^n)\) be the Schwartz space with the classical topology and denote by \(S'(\mathbb{R}^n)\) its toplogical dual. We also let \[ S_{\infty }(\mathbb{R}^n)=\left\{ \varphi \in S(\mathbb{R}^n):\int_{\mathbb{R}^n}\varphi (x) x^{\gamma }dx=0,\text{for all }\gamma \in(\mathbb{N}\cup \{0\}) ^n\right\} \] and consider \(S_{\infty }(\mathbb{R}^n)\) as a subspace of \(S(\mathbb{R}^n),\) including the topology. Let \(S_{\infty }'(\mathbb{R}^n)\) be its topological dual. For any \(\varphi \in S(\mathbb{R}^n)\), we use \(\hat{\varphi}\) to denote its Fourier transform. Let \(\varphi _{j}(x) =2^{jn}\varphi (2^{j}x) \) for all \(j\in\mathbb{Z}\) and \(x\in\mathbb{R}^n\). For \(j\in\mathbb{Z}\) and \(k\in\mathbb{Z}^n,\) denote by \(Q_{jk}\) the dyadic cube \(Q_{jk}=2^{-j}([0,1)^n+k)\). Let \[ Q\left(\mathbb{R}^n\right) =\left\{ Q_{jk}:j\in\mathbb{Z}\text{ and }k\in\mathbb{Z}^n\right\} ,\;Q_{j}\left(\mathbb{R}^n\right) =\left\{\mathbb{Q}\in Q\left(\mathbb{R}^n\right) :l(\mathbb Q) =2^{-j}\right\}, \] and \(j_{\mathbb{Q}}=-\log _{2}l(\mathbb Q) \) for all \(\mathbb{Q}\in Q(\mathbb{R}^n),\) where \(l(\mathbb Q)\) is its side length. Let \(\varphi \in S(\mathbb{R}^n)\) such that \[ \text{supp\,}\hat{\varphi}\subset \left\{\xi \in\mathbb{R}^n:\tfrac{1}{2}\leq \left| \xi \right| \leq 2\right\} \text{ and }\left| \hat{\varphi}(\xi)\right| \geq C\text{ if }\tfrac{3}{5}\leq \left| \xi \right| \leq \tfrac{5}{3}, \tag{1} \] where \(C\) is a positive constant independent of \(\xi\). Let \(q\in (0,\infty ]\) and \(\tau \in [ 0,\infty )\). Denote by \(\ell^{q}( L_{\tau }^{p}(\mathbb{R}^n) ) \) with \(p\in (0,\infty ]\) the set of all sequences \(G=\{ g_{j}\} _{j\in\mathbb{Z}}\) of measurable functions on \(\mathbb{R}^n\) such that \[ \left\| G\right\| _{_{\ell ^{q}\left( L_{\tau }^{p}\left(\mathbb{R}^n\right)\right) }}=\sup_{P\in Q\left(\mathbb{R}^n\right) }\frac{1}{\left| P\right| ^{\tau }}\left\{\sum_{j=j_{P}}^{\infty }\int_{P}\left( \left| g_{j}\left( x\right) \right| ^{p}dx\right) ^{\frac{q}{p}}\right\} ^{\frac{1}{q}}<\infty. \] Similarly, for \(p\in (0,\infty )\) the space \(L_{\tau }^{p}( \ell ^{q}(\mathbb{R}^n)) \) is defined to be the set of sequences \(G=\{g_{j}\} _{j\in\mathbb{Z}}\) of measurable functions on \(\mathbb{R}^n\) such that \[ \left\| G\right\| _{L_{\tau }^{p}\left( \ell ^{q}\left(\mathbb{R}^n\right)\right) }=\sup_{P\in Q\left(\mathbb{R}^n\right) }\frac{1}{\left| P\right| ^{\tau }}\left\{\int_{P}\left( \sum_{j=j_{P}}^{\infty }\left| g_{j}(x)\right| ^{p}\right) ^{\frac{q}{p}}dx\right\} ^{\frac{1}{q}}<\infty. \] Assume that \(s\in\mathbb{R}\), \(t\in [ 0,\infty )\), \(q\in (0,\infty ]\) and \(\varphi \in S(\mathbb{R}^n)\) satisfy (1). The Besov-type space \(\dot{B}_{p,q}^{s,\tau }(\mathbb{R}^n) \) with \(p\in (0,\infty ]\) is defined to be the set of all \(f\in S_{\infty }'(\mathbb{R}^n)\) such that \(\| f\| _{\dot{B}_{p,q}^{s,\tau}(\mathbb{R}^n) }=\|\{ 2^{js}( \varphi _{j}\ast f)\} _{j\in\mathbb{Z}}\| _{\ell ^{q}( L_{\tau }^{p}(\mathbb{R}^n)) }<\infty\). Also, the Triebel-Lizorkin-type space \(\dot{F}_{p,q}^{s,\tau }(\mathbb{R}^n) \) with \(p\in (0,\infty )\) is defined to be the set of all \(f\in S_{\infty }'(\mathbb{R}^n)\) such that \(\| f\| _{\dot{F}_{p,q}^{s,\tau }(\mathbb{R}^n) }=\|\{ 2^{js}( \varphi _{j}\ast f)\} _{j\in\mathbb{Z}}\| _{L_{\tau }^{p}( \ell ^{q}(\mathbb{R}^n)) }<\infty\). Let \(\varphi \in S(\mathbb{R}^n)\) satisfy \((1)\) and \(f\in S_{\infty}'(\mathbb{R}^n)\). For all \(j\in\mathbb{Z}\), \(a\in (0,\infty )\), \(P\in Q(\mathbb{R}^n) \) and \(x\in\mathbb{R}^n,\) define \[ \varphi _{j}^{\ast ,a}f(x)=\sup_{y\in\mathbb{R}^n}\,\frac{\left| \varphi _{j}\ast f\left( y\right) \right| }{\left(1+2^{j}\left| x-y\right| \right) ^{a}},\quad \varphi _{j,P}^{\ast,a}f(x)=\sup_{y\in P}\,\frac{\left| \varphi _{j}\ast f\left( y\right) \right| }{\left( 1+2^{j}\left| x-y\right|\right) ^{a}}. \tag{2} \] In this paper, in Theorem 1.1, the authors use the maximal functions in (2) to characterize the spaces \(\dot{B}_{p,q}^{s,\tau }(\mathbb{R}^n)\) and \(\dot{F}_{p,q}^{s,\tau }(\mathbb{R}^n) \) for all admissible indices \(s,\tau ,p\) and \(q\). For an integer \(\beta \geq 1\), let \(S_{-1}(\mathbb{R}^n)=S(\mathbb{R}^n)\) and \(S_{\beta }(\mathbb{R}^n)\) be the set of all \(\phi \in S(\mathbb{R}^n)\) satisfying \(\int_{\mathbb{R}^n}\phi (x)x^{\gamma }\,dx=0\) for all \(|\gamma| \leq \beta\) when \(\beta \geq 0\). Consider \(S_{\beta }(\mathbb{R}^n)\) as a subspace of \(S(\mathbb{R}^n),\) including the topology and \(S_{\beta }'(\mathbb{R}^n)\) its topological dual. Also, let \(\varepsilon \in (0,\infty )\) and \(k\in S_{\beta }(\mathbb{R}^n)\) satisfy the following Tauberian condition: \[ |\hat k(\xi)| >0\text{ on }\left\{ \xi \in\mathbb{R}^n:\tfrac{\varepsilon }{2}<\left| \xi \right| <2\varepsilon \right\}. \tag{3} \] For \(f\in S_{\beta }'(\mathbb{R}^n)\), define \(k_{j}^{\ast ,a}f(x)\) and \(k_{j,P}^{\ast ,a}f(x)\) as in (2). \(k_{j}\ast f\) is called local mean. In Theorem 1.2 of this paper, the authors establish the local mean characterizations of \(\dot{B}_{p,q}^{s,\tau }(\mathbb{R}^n)\) and \(\dot{F}_{p,q}^{s,\tau }(\mathbb{R}^n)\). For any \(p,q\in (0,\infty ],\) define \(p\wedge q=\min\{ p,q\} \) and \(p\vee q=\max\{ p,q\}\). Set \(\mathbb{R}_{+}^{n+1}=\mathbb{R}^n\times (0,\infty )\). For any measurable function \(w\) on \(\mathbb{R}_{+}^{n+1}\) and \(x\in\mathbb{R}^n,\) define \(N_{w}(x)=\sup_{|y-x|<t}|w(y,t)|\). Assume that \(p\in (1,\infty)\) and \(\tau \in [ 0,\infty )\). Denote by \(\widetilde{\ell ^{q}( L_{\tau }^{p}(\mathbb{R}^n)) }\) with \(q\in [ 1,\infty )\) the set of all sequences \(G=\{g_j\} _{j\in\mathbb{Z}}\) of measurable functions on \(\mathbb{R}^n\) such that \[ \left\| G\right\| _{\widetilde{\ell ^{q}\left( L_{\tau}^{p}\left(\mathbb{R}^n\right) \right) }}=\inf_{w}\left\{ \sum_{j\in\mathbb{Z}}\left( \int_{\mathbb{R}^n}| g_{j}(x)| ^{p}\left[ w(x,2^{-j}) \right] ^{-p}dx\right) ^{\frac{q}{p}}\right\} ^{\frac{1}{q}}<\infty \] and by \(\widetilde{L_{\tau }^{p}( \ell ^{q}(\mathbb{R}^n)) }\) with \(q\in (1,\infty )\) the set of all sequences \(G=\{ g_j\} _{j\in\mathbb{Z}}\) of measurable functions on \(\mathbb{R}^n\) such that \[ \left\| G\right\| _{\widetilde{L_{\tau }^{p}\left( \ell ^{q}\left(\mathbb{R}^n\right)\right) }}=\inf_{w}\left\{ \int_{\mathbb{R}^n}\left( \sum_{j\in\mathbb{Z}}\left( \left| g_{j}(x)\right| ^{q}\left[ w\left( x,2^{-j}\right) \right] ^{-q}dx\right) \right) ^{\frac{p}{q}}\right\} ^{\frac{1}{p}}<\infty , \] where the infimum is taken over all non-negative Borel measurable functions on \(\mathbb{R}_{+}^{n+1}\) satisfying \[ \int_{\mathbb{R}^n}\left[ N_{w}(x)\right] ^{\left( p\vee q\right) '}dH^{n\tau \left( p\vee q\right) '}(x) <1 \] and with the restriction that for any \(j\in\mathbb{Z},\) \(w( .,2^{-j}) \) is allowed to vanish only where \(g_{j}\) vanishes, where the integral is the Choquet integral and \(( p\vee q) '\) denotes the conjugate index of \(p\vee q.\) Let \(s\in \mathbb{R},\) \(p\in (1,\infty )\) and \(\varphi \in S(\mathbb{R}^n)\) satisfy \((1)\). The Besov-Hausdorff space \(\dot{BH}_{p,q}^{s,\tau }(\mathbb{R}^n)\) with \(q\in [ 1,\infty )\) and \(\tau \in[ 0,\frac{1}{( p\vee q '}] \) is defined to be the set of all \(f\in S_{\infty }'(\mathbb{R}^n)\) such that \(\| f\| _{\dot{BH}_{p,q}^{s,\tau}(\mathbb{R}^nt) }=\| \{ 2^{js}( \varphi _{j}\ast f)\} _{j\in\mathbb{Z}}\| _{\widetilde{\ell ^{q}( L_{\tau }^{p}(\mathbb{R}^n) ) }}<\infty\). Also, the Triebel-Lizorkin-Hausdorff space \(\dot{FH}_{p,q}^{s,\tau }(\mathbb{R}^n) \) with \(q\in (1,\infty )\) and \(\tau \in[ 0,\frac{1}{( p\vee q) '}] \) is defined to be the set of all \(f\in S_{\infty }'(\mathbb{R}^n)\) such that \(\| f\| _{\dot{FH}_{p,q}^{s,\tau}(\mathbb{R}^n}=\|\{ 2^{js}( \varphi _{j}\ast f)\} _{j\in\mathbb{Z}}\| _{\widetilde{L_{\tau }^{p}( \ell ^{q}(\mathbb{R}^n)) }}<\infty\). In Theorem 1.3 and Theorem 1.4, the authors establish maximal function characterizations of \(\dot{BH}_{p,q}^{s,\tau }(\mathbb{R}^n)\) and \(\dot{FH}_{p,q}^{s,\tau }(\mathbb{R}^n) \) for all admissible indices \(s,\tau ,p\) and \(q\). At the end of the paper, the authors give an application of Theorems 1.1 and 1.3. They establish a Fourier multiplier theorem of the spaces \(\dot{B}_{p,q}^{s,\tau }(\mathbb{R}^n) \), \(\dot{F}_{p,q}^{s,\tau }(\mathbb{R}^n) ,\dot{BH}_{p,q}^{s,\tau }(\mathbb{R}^n)\) and \(\dot{FH}_{p,q}^{s,\tau }(\mathbb{R}^n) \) in Theorem 4.1.