Gaussian \(\mathrm{JN}_p\) spaces (Q6948641)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 8074505
Language Label Description Also known as
default for all languages
No label defined
    English
    Gaussian \(\mathrm{JN}_p\) spaces
    scientific article; zbMATH DE number 8074505

      Statements

      Gaussian \(\mathrm{JN}_p\) spaces (English)
      0 references
      0 references
      0 references
      0 references
      5 August 2025
      0 references
      For \(d\gamma=\pi^{-d/2} e^{-|x|^2}\, dx\), \(m(x)=\min\{1,1/|x|\}\) (\(x\neq 0\)) and \(a>0\), \(\mathcal{B}_a\) denotes the set of balls \(B=B(c_B,r_B)\) with center \(c_B\) and radius \(r_N\leq a m(c_B)\) and \(\mathcal{Q}_a\) consists of those cubes \(Q\subset\mathbb{R}^d\) with sidelength \(\ell_Q\leq a m(c_Q)\). For \(1<p<\infty\), \(f\in L^1(\mathbb{R}^d,\gamma)\) is said to lie in \(\mathrm{JN}_p^{\mathcal{Q}_a}(\mathbb{R}^d,\gamma)\) if\N\[\N{K}_p^{\mathcal{Q}_a}(f)=\sup\left(\sum_i \gamma(Q_i)\left(\frac{1}{\gamma(Q_i)}\int_{Q_i} |f-f_{Q_i}|\, d\gamma\right)^p\right)^{1/p}<\infty\N\]\Nwhere the supremum is taken over countable collections of pairwise disjoint cubes in \(\mathcal{Q}_a\) and \(f_Q\) is the mean value of \(f\) on \(Q\) with respect to \(d\gamma\). Then\N\[\N\|f\|_{\mathrm{JN}_p^{\mathcal{Q}_a}(\mathbb{R}^d,\gamma)}=\|f\|_{L^1(\mathbb{R}^d,\gamma)}+{K}_p^{\mathcal{Q}_a}(f)\N\]\Ndefines a norm on \(\mathrm{JN}_p^{\mathcal{Q}_a}(\mathbb{R}^d,\gamma)\). The following John--Nirenberg type inequality is proved (Theorem.~1.1): Let \(a>0\) and \(1<p<\infty\). There exists \(C>0\) such that, for all \(Q\in\mathcal{Q}_a\) and \(f\in \mathrm{JN}_p^{\mathcal{Q}_a}(\mathbb{R}^d,\gamma)\),\N\[\N\gamma (\{x\in Q: |f-f_Q|>\sigma\})\leq C\left(\frac{{K}_p^{\mathcal{Q}_a}(f)}{\sigma}\right)^p \, .\N\]\N\NThe main result (Theorem.~1.2) characterizes \(\mathrm{JN}_p^{\mathcal{Q}_a}(\mathbb{R}^d,\gamma)\) as the dual of an atomic-type of Hardy space associated with \(\mathcal{Q}_a\) and \(\gamma\). One says that \(b\in \mathcal{A}(q,a,Q)\) if \(\mathrm{supp}\, b \subset Q\in\mathcal{Q}_a\) and \(b\in L^q_0(Q,\gamma)\) where the subscript indicates that \(\int_Q f\, d\gamma =0\). One then says that \(g\) is a \((p,q,a)\)-polymer if \(g=\sum_{j=1}^\infty b_j\) with \(b_j\in \mathcal{A}(q,a, Q_j)\) (with \(\{Q_j\}\subset\mathcal{Q}_a\) pairwise disjoint) and\N\[\N\|g\|_{(p,q,a)}=\inf \sum \lambda_j<\infty \quad\mathrm{where}\quad \lambda_j=\gamma(Q_j)\left(\frac{1}{\gamma(Q_j)} \int_{Q_j} |b_j|^q\, d\gamma\right)^{p/q}\, .\N\]\NOne then says that a measurable function \(g\) belongs to \(H_{p,q,a}(\mathbb{R}^d,\gamma)\) if \(g=c_0+\sum_{i=1}^\infty g_i\) where \(c_0\in\mathbb{C}\), each \(g_i\) is a \((p,q,a)\)-polymer, and \(\sum_{i=1}^\infty \|g_i\|_{(p,q,a)}<\infty\). Also, \(\|g\|_{H_{p,q,a}(\mathbb{R}^d,\gamma)}=|c_0|+\inf \sum_{i=1}^\infty \|g_i\|_{(p,q,a)}\) with the infimum taken over polymer decompositions of \(g\).\N\NGiven \(f\in \mathrm{JN}_p^{\mathcal{Q}_a}(\mathbb{R}^d,\gamma)\) one sets \(\Lambda_fg=\lim_{N\to\infty} \int_{\mathbb{R}^d} f_N g\, d\gamma\) where \(f_N=\mathrm{sgn}\, f \min\{|f|,N\}\). The main theorem characterizes \( \mathrm{JN}_p^{\mathcal{Q}_a}(\mathbb{R}^d,\gamma)\) as the dual of \(H_{p,q,a}(\mathbb{R}^d,\gamma)\). Specifically, there is a fixed \(C>0\) such that if \(f\in \mathrm{JN}_p^{\mathcal{Q}_a}(\mathbb{R}^d,\gamma)\) then \(\Lambda_f\in (H_{p,q,a}(\mathbb{R}^d,\gamma))'\) and\N\[\N\|\Lambda_f\|_{(H_{p,q,a}(\mathbb{R}^d,\gamma))^\prime}\leq C\|f\|_{ \mathrm{JN}_p^{\mathcal{Q}_a}(\mathbb{R}^d,\gamma)}\, .\N\]\NConversely, if \(\Lambda \in (H_{p,q,a}(\mathbb{R}^d,\gamma))^\prime\) then there is an \(f\in \mathrm{JN}_p^{\mathcal{Q}_a}(\mathbb{R}^d,\gamma)\) such that \(\Lambda=\Lambda_f\) and\N\[\N\|f\|_{ \mathrm{JN}_p^{\mathcal{Q}_a}(\mathbb{R}^d,\gamma)}\leq C \|\Lambda\|_{(H_{p,q,a}(\mathbb{R}^d,\gamma))'}\, .\N\]\NAmong properties and tools, it is shown that (algebraically and topologically), \(\mathrm{ JN}_p^{\mathcal{Q}_a}(\mathbb{R}^d,\gamma)\) does not depend on \(a\) and that \(\lim_{p\to\infty} \|f\|_{ \mathrm{JN}_p^{\mathcal{Q}_a}(\mathbb{R}^d,\gamma)}= \|f\|_{\mathrm{BMO}^{\mathcal{Q}_a}(\mathbb{R}^d,\gamma)}\) where the BMO space is suitably defined for \(\gamma\) and \(\mathcal{Q}_a\). The John-Nirenberg inequality (Theorem.~1.1) is then proved. Duality (Theorem.~1.2) requires some preliminary identification of suitable dense subspaces, independence of \(H_{p,q,a}(\mathbb{R}^d,\gamma)\) with respect to \(a\), covering properties, and technical estimates for polymer decompositions that ultimately allow for representation of functionals on \(H_{p,q,a}(\mathbb{R}^d,\gamma)\) by specific elements of \(\mathrm{JN}_p^{\mathcal{Q}_a}(\mathbb{R}^d,\gamma)\).
      0 references
      0 references
      John-Nirenberg spaces
      0 references
      Gaussian measure
      0 references
      Hardy spaces
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references