Collectively compact composition operator sequences between Bergman and Hardy spaces (Q705367)

Assume that \(H_1, \, H_2\) are Hilbert spaces and that \(\{T_n\}\) is a sequence of uniformly bounded linear operators from \(H_1\) to \(H_2\). According to G. T. Zhu, M. Z.~Yang, S. K.~Li, X. M.~Chen and W. J.~Li, the sequence \(\{T_n\}\) is said to be a collectively compact operator sequence (CCOS) if \(\lim_{N \to \infty} \inf \{\varepsilon > 0:\) for every sequence \(\{x_k\}\) converging weakly to \(0\) with \(\| x_k\| _{H_1} \leq 1\), there is a \(K\) such that \(\| T_n x_k\| _{H_2} < \varepsilon\) when \(n > N\) and \(k > K\} = 0\). The main result of the present paper, which extends several results by J. H.~Shapiro that characterize the compactness of the single composition operator \(C_\varphi\) (where \(\varphi :D \to D\) is an analytic self-map of the open unit disk \(D\) in the complex plane, and \(C_\varphi (f) := f \circ \varphi\) for every holomorphic function \(\varphi\) on \(D\)) on the Hardy space \(H^2\) and the Bergman space \(A^2\), is the following Theorem. Let \(\{\varphi_n\}\) be a sequence of analytic self-maps of \(D\), and consider for every \(\gamma > -1\) the weighted Bergman space \(A^2_\gamma\), that is, the Hilbert space consisting of those analytic functions \(f\) on \(D\) for which \(\| f\| := (\int \int_D | f(z)| ^2 (\log {1 \over | z| })^\gamma \, {dxdy \over \pi})^{1/2} < \infty\) (note that \(A^2_0 = A^2\)). Define also \(A^2_{-1} := H^2\). Assume that \(\alpha , \, \beta \geq -1\) and that \(\{C_{\varphi_n}\}\) is a sequence of uniformly bounded composition operators from \(A^2_\alpha\) to \(A^2_\beta\). Then \(\{C_{\varphi_n}\}\) is a CCOS if and only if \[ \lim_{| w| \to 1 \atop n \to \infty} {N_{\varphi_n, \beta + 2}(w) \over (\log {1 \over | w| })^{\alpha + 2}} = 0. \] Here, \(N_{\varphi, \alpha}\) (for \(\varphi\) an analytic self-map of \(D\) and \(\gamma \geq 0\)) is the Nevanlinna generalized counting function, defined as \(N_{\varphi, \alpha} (w) = \sum_{z \in \varphi^{-1}(w)} (\log {1 \over | z| })^\gamma\) \((w \in D \setminus \{\varphi (0)\})\), where \(\varphi^{-1}(w)\) denotes the set of \(\varphi\)-preimages of \(w\) counting the multiplicity, and \(N_{\varphi, \alpha} (w) = 0\) if \(w \notin \varphi (D)\).