Bergman projection induced by kernel with integral representation (Q2330785): Difference between revisions

Let \(\hat{\mathcal{D}}\) denote the set of positive Borel measures \(\omega\) on [0,1) such that \(\hat{\omega}(r)=\int_r^1 d\omega (r)\leq C\hat{\omega}(\frac{1+r}{2})\) for some \(C=C(\omega)>0\). For \(0<p<\infty\) the weighted Bergman space \(A^p_\omega\) consists of analytic functions \(f\) in the unit disk \(\mathbb{D}\) such that \(\|f\|^p_{A^p_\omega}=\int_\mathbb{D}|f(z)|^p d(\omega \otimes m)(z) <\infty\), where \(d(\omega \otimes m)(re^{i\theta} )=rd\omega(r)d\theta\). Since the point-evaluation \(L_z(f)=f(z)\) is a bounded linear functional in \(A^2_\omega\) by an earlier result of \textit{J. A. Peláez} [``Small weighted Bergman spaces'', in: Proceedings of the Summer School in Complex and Harmonic Analysis, and Related Topics. Joensuu: University of Eastern Finland, Faculty of Science and Forestry. 29--98 (2016)], there exist unique reproducing kernels \(B_z^\omega \in A^2_\omega\) with \(\|L_z\|=\|B_z^\omega\|_{A^2_\omega}\), such that \[ f(z)= \int_\mathbb{D}f(\zeta) \overline{B^\omega_z(\zeta)} d(\omega \otimes m)(\zeta). \] The Bergman projection \(P_\omega (f)(z)= \int_\mathbb{D}f(\zeta) \overline{B^\omega_z(\zeta)} d(\omega \otimes m)(\zeta)\) is an orthogonal projection from \(L_\omega^2\) to \(A_\omega^2\), and it is closely related to the maximal Bergman projection \(P^+_\omega (f)(z)= \int_\mathbb{D}f(\zeta) | B^\omega_z(\zeta) |d(\omega \otimes m)(\zeta)\). For a positive Borel measure \(\omega\) on [0,1), a positive \((\omega \otimes m)\)-integrable function is called an \(\omega\)-weight. For \(0<p<\infty\) and an \(\omega\)-weight \(\upsilon\), the Lebesgue space \(L^p_\omega (\upsilon)\) consists of \(f\) such that \(\|f\|^p_{L^p_\omega (\upsilon)}=\int_\mathbb{D}|f(z)|^p \upsilon(z)d(\omega \otimes m)(z) <\infty\). The authors derive a number of interesting results (too numerous to mention all of them) about the Bergman projections and the maximal Bergman projections induced by kernels with integral representation. For a finite positive measure \(\upsilon\) supported on [0,1] such that \(\int_0^1 \frac{d\upsilon(r)}{1-r}\) diverges, they show that there exists \(\omega \in \hat{\mathcal{D}}\) such that \[ B_z^\omega(\zeta)=\frac{1}{1-\overline{z}\zeta}\int_0^1 \frac{d\upsilon(r)}{1-r\overline{z}\zeta}, \] \(z,\zeta \in \mathbb{D}\). Then they extend the classical Bekollé-Bonami's results to \(\omega \in \mathcal{R}\) that induce kernels with the representation \[ B_z^\omega(\zeta)=\frac{1}{(1-\overline{z}\zeta)^\gamma}\int_0^1 \frac{d\upsilon(r)}{1-r\overline{z}\zeta}, \] \(z,\zeta \in \mathbb{D}\), \(1\leq \gamma\). For \(1<p<\infty\) and \(\omega \in \hat{\mathcal{D}}\), an \(\omega\)-weight \(\upsilon\) belongs to \(B_{p,\omega}\) if \[ B_{p,\omega}(\upsilon)=\sup_S\frac{(\upsilon \omega)(S)}{\omega (S)}\Big[\frac{(\upsilon^{-p'/p} \omega)(S)}{\omega(S)}\Big]^{p/p'}<\infty, \] where \(S\) is a Carleson square in \(\mathbb{D}\). The authors prove that the statements \(P^+_\omega :L^p_\omega (\upsilon)\rightarrow L^p_\omega (\upsilon)\) is bounded, \(P_\omega :L^p_\omega (\upsilon)\rightarrow L^p_\omega (\upsilon)\) is bounded, \(P_\omega :L^p_\omega (\upsilon)\rightarrow L^{p,\infty}_\omega (\upsilon)\) is bounded, and \(\upsilon \in B_{p,\omega}\) are equivalent, and that \(\| P^+_\omega\|_{L^p_\omega (\upsilon)\rightarrow L^p_\omega (\upsilon)} \asymp B_{p,\omega} (\upsilon)^{\max [1, \frac{1}{p-1}]}\). They also characterize the weak \((1,1)\)-inequality in terms of Bekollé-Bonami-type conditions and they discuss the two-weight inequality for \(P^+_\omega :L^p_\omega (\upsilon)\rightarrow L^p_\omega (\upsilon)\) in terms of Sawyer-testing conditions.