On generalized Besov and Triebel-Lizorkin spaces of regular distributions (Q393055): Difference between revisions

A sequence \(\sigma = (\sigma_j )^\infty_{j=0}\) of positive numbers is called admissible if \(\sigma_j \sim \sigma_{j+1}\) (equivalence at the expense of two positive constants which are independent of \(j\)). Let \(\{ \phi_j \}^\infty_{j=0}\) be the usual dyadic resolution of unity in \(\mathbb R^n\). Then \(B^\sigma_{p,q} (\mathbb R^n)\) with \(0<p,q \leq \infty\) collects all \(f\in S'(\mathbb R^n)\) such that \[ \Big( \sum^\infty_{j=0} \sigma_j \big\| (\phi_j \hat{f} )^\vee | L_p (\mathbb R^n) \big\|^q \Big)^{1/q} \] is finite. Similarly, \(F^\sigma_{p,q} (\mathbb R^n)\) with \(L_p (\ell_q)\) in place of \(\ell_q (L_p)\), \(0<p<\infty\), \(0<q\leq \infty\). This generalizes the classical spaces \(A^s_{p,q} (\mathbb R^n)\), where \(\sigma_j =2^{js}\) with \(s\in \mathbb R\). The authors also deal with the more general spaces \(B^{\sigma,N}_{p,q} (\mathbb R^n)\) and \(F^{\sigma,N}_{p,q} (\mathbb R^n)\) based on a resolution of unity \(\{ \phi_j \}\) with \[ \text{supp }\phi_j \subset \big\{ \xi \in \mathbb R^n: \quad N_{j-j_0} \leq |\xi| \leq N_{j + j_0} \big\}, \] where \(N = \{ N_j \}\), \[ 0< \lambda_0 N_j \leq N_{j+1} \leq \lambda_1 N_j, \qquad 1<\lambda_0 \leq \lambda_1, \] with some natural number \(j_0\). But the corresponding spaces \(A^{\sigma,N}_{p,q} (\mathbb R^n)\), \(A\in \{B,F\}\), can be reduced to some \(A^{\tilde{\sigma}}_{p,q} (\mathbb R^n)\) (called standardization). It is the main aim of the paper (Theorems 4.3 and 4.4) to characterize (by necessary and sufficient conditions) those spaces \(A^{\sigma,N}_{p,q} (\mathbb R^n)\) which consist entirely of regular distributions, i.e., \[ A^{\sigma,N}_{p,q} (\mathbb R^n) \subset L^{\mathrm{loc}}_1 (\mathbb R^n). \]