Continuity envelopes and sharp embeddings in spaces of generalized smoothness (Q2426481)

Let \({\mathbb N}_0=\mathbb N \cup \{0\}\) and \(\mathcal{N}=\{N_j\}_{j\in{\mathbb N}_0}\) be an admissible sequence, which means that there exist two positive constants \(d_0\) and \(d_1\) such that \(d_0\gamma_j\leq\gamma_{j+1}\leq d_1\gamma_j\) for \(j\in{\mathbb N}_0\). Assume that \(\underline{N}_1\equiv\inf_{k\geq0}\frac{\gamma_{1+k}}{\gamma_k}>1\). Then there exists a number \(l_0\in{\mathbb N}_0\) such that \(N_k\geq2N_j\) for any \(k,\;j\) such that \(k\geq j+l_0\). The associated covering \(\Omega^N=\{\Omega_j^N\}_{j\in{\mathbb N}_0}\) of \({\mathbb R}^n\) is defined by setting \(\Omega_j^N\equiv\{\xi\in{\mathbb R}^n: |\xi|\leq N_{j+l_0}\},\) \(j=0, \dots, l_0-1\), and \(\Omega_j^N\equiv\{\xi\in{\mathbb R}^n:\;N_{j-l_0}\leq|\xi|\leq N_{j+l_0}\},\) \(j\geq l_0\). Let \(\varphi^N\equiv\{\varphi_j^N\}_{j\in{\mathbb N}_0}\) be a (generalized) partition of unity subordinate to \(\Omega^N\), which means that (i) \(\varphi_j^N\in C_0^\infty({\mathbb R}^n)\) and \(\varphi_j^N(\xi)\geq0\) if \(\xi\in{\mathbb R}^n\) for any \(j\in{\mathbb N}_0\); (ii) \(\text{supp}\varphi_j^N\subset\Omega_j^N\) for any \(j\in{\mathbb N}_0\); (iii) for any \(\alpha\in{\mathbb N}^n_0\), there exists a positive constant \(C_\alpha\) such that for any \(j\in{\mathbb N}_0\), \(|D^\alpha\varphi_j^N(\xi)|\leq C_\alpha(1+|\xi|^2)^{-|\alpha|/2}\) for any \(\xi\in{\mathbb R}^n\) and (iv) there exists a positive constant \(C_\varphi\) such that \(0<\sum_{j=0}^\infty\varphi_j^N(\xi)=C_\varphi<\infty\) for any \(\xi\in{\mathbb R}^n\). Let \(\sigma=\{\sigma_j\}_{j\in{\mathbb N}_0}\) be an another admissible sequence. If \(p, q\in(0,\infty]\), the Besov space of generalized smoothness \(B^{\sigma,N}_{p,q}\) is the set of all tempered distributions \(f\) such that the quasi-norm \[ \|f\|_{B^{\sigma,N}_{p,q}}=\left\{\sum_{j=0}^\infty \|\sigma_j\varphi_j^N(D)f\|_{L^p}^q\right\}^{1/q}<\infty. \] If \(p\in(0, \infty)\) and \(q\in(0,\infty]\), the Triebel-Lizorkin space of generalized smoothness \(F^{\sigma,N}_{p,q}\) is the set of all tempered distributions \(f\) such that the quasi-norm \[ \|f\|_{F^{\sigma,N}_{p,q}}=\left\|\left\{\sum_{j=0}^N |\sigma_j\varphi_j^N(D)f|^q\right\}^{1/q}\right\|_{L^p}<\infty. \] Recall that, roughly speaking, a continuity envelope \({\mathfrak C}_c(X)\) of a function space \(X\) consists of a so-called continuity envelope function \[ \mathcal {E}_c^X(t)\sim\sup_{\|f|X\|\leq1}\frac{\omega(f,\,t)}{t},\;t>0, \] together with some `finite index' \(u_X\), where \(\omega(f,\,t)\) stands for the modulus of continuity. Let \(\Lambda\) be a certain function satisfying \(\Lambda(t)\sim\sigma_j\) for \(z\in[N_j,N_{j+1}]\) and \(j\in{\mathbb N}_0\); then the authors prove that \[ \mathcal {E}_c^{B^{\sigma,N}_{p,q}}(t)\sim \left(\int_t^{N_0^{-1}}\Lambda(y^{-1})^{-q'} y^{-(\frac{n}{p}+1)q'}\frac{dy}{y}\right)^{1/q'}, \] if \(q\in(1,\infty]\), otherwise, for \(q\in(0,1]\), appropriately modified by \[ \mathcal {E}_c^{B^{\sigma,N}_{p,q}}(t)\sim\sup_{y\in[t,N_0^{-1}]} \Lambda(y^{-1})^{-1}y^{-(\frac{n}{p}+1)}, \] where \(1/q+1/q'=1\). There exists a parallel result for \(F\)-spaces. In fact, the authors give even more precise characterizations leading to the concept of so-called continuity envelopes, where the continuity envelope function \(\mathcal {E}_c^X\) is complemented by the fine index \(u_X\). Moreover, if the lower Boyd index \(\underline{s}(\sigma N^{-n/p})>0\), the authors obtain the sharp embedding result that \(A^{\sigma,N}_{p,q}({\mathbb R}^n)\hookrightarrow \text{Lip}^1({\mathbb R}^n)\) if and only if \(\sigma^{-1}N^{n/p+1}\in l_{u'_A}\), where \(A\) denotes \(B\) or \(F\), \(u_A=q\) when \(A=B\), and \(u_A=p\) when \(A=F\), and \(\frac{1}{u'_A}=\max(1-\frac{1}{u_A},0)\). The authors also prove a criterion for the embedding \(A^{\sigma,N}_{p_1,q_1}({\mathbb R}^n)\hookrightarrow A^{\tau,N}_{p_2,q_2}({\mathbb R}^n)\) in the limiting case \(\underline{s}(\sigma N^{-n/{p_1}})=\overline{s}(\tau N^{-n/{p_2}})\). Finally, these envelope results are used to obtain some Hardy type inequalities.