Multipliers on vector spaces of holomorphic functions

From MaRDI portal





Let \(H(G)\) be the space of holomorphic functions on a domain \(G\subset{\mathbb C}\) containing zero. The Hadamard product of two power series \(f(z)=\sum_{n=0}^\infty a_nz^n\) and \(g(z)=\sum_{n=0}^\infty b_nz^n\) is defined by \(f*g(z)=\sum_{n=0}^\infty a_nb_nz^n\). Let \(G_1\) and \(G_2\) be two domains in \({\mathbb C}\) containing zero. A power series \(g(z)=\sum_{n=0}^\infty b_nz^n\) is a coefficient multiplier if \(g*f\in H(G_2)\) for all \(f\in H(G_1)\), i.e., if \(T_g(f):=g*f\) defines a linear mapping \(T_g: H(G_1)\to H(G_2)\). For the case \(G_1=G_2=G\); \(M(H(G))\) denotes the algebra (with respect to composition) of all coefficient multipliers. A domain \(G\) is admissible if the set \(H(G)\) is an algebra with respect to the Hadamard product. The authors consider the following questions: is it possible to identify the coefficient multiplier algebra \(M(H(G))\) with a certain vector space of holomorphic functions? What does it mean that two coefficient multiplier algebras \(M(H(G_1))\) and \(M(H(G_2))\) are isomorphic? The central object of investigation is the domain \(\widehat{G}\) introduced by \textit{N. U. Arakelyan} [see Math. USSR, Sb. 52, 21-39 (1985; Zbl 0571.30003) or the original article in Russian Mat. Sb., Nov. Ser. 124(166), No. 1(5), 24-44 (1984; Zbl 0557.30004)]: \[ \widehat{G}:=\bigcup_{w\in G^c}{1\over w}G. \] The authors use the number \(k_G\) in the next definition as a characteristic of the domain \(G\). Let \(G\) be a domain containing \(0\). For \(k\in{\mathbb N}\) \(A_k\) denotes the set of all \(k\)-th roots of unity. If there exists a largest number \(k\in{\mathbb N}\) such that \(\xi w\in G^c\) for all \(\xi\in A_k\), \(w\in G^c\), then this number is denoted by \(k_G\). The main result of the article states that the isomorphy of multiplier algebras \(M(H(G_1))\) and \(M(H(G_2))\) implies that \(\widehat{G_1}\) is necessarily equal to \(\widehat{G_2}\). The authors describe all isomorphisms between two coefficient multiplier algebras in the theorem 3.4: Let \(G_1\), \(G_2\) be domains containing 0 and different from \({\mathbb D}_r=\{z\in{\mathbb C}: |z|<r\}\) for all \(r>0\). Let \(\Psi: M(H(G_1))\to M(H(G_2))\) be an isomorphism. Then \(k:=k_{G_1}=k_{G_2}\) and there exist \(n_0\in{\mathbb N}_0\) and \(b_0,\dots,b_{k-1}\in{\mathbb Z}\) such that \(\psi(kn+j)=kn+b_j\) for all \(nk+j\geq n_0\) and for all \(j=0,\dots,k-1\) where \(\psi: {\mathbb N}_0\to{\mathbb N}_0\) is given by the formula \(\Psi(L_{z^n})=L_{z^{\psi(n)}}\) for all \(n\in{\mathbb N}_0\) (here \(L_{z^n}(f)=z^n*f\)). As a corollary the next statement holds true: if \(\Phi: M(H(G_1))\to M(H(G_2))\) is an isomorphism, then \(\widehat{G_1}=\widehat{G_2}\). The authors show that if \(G_1\) and \(G_2\) are admissible domains such that \(H(G_1)\) and \(H(G_2)\) are isomorphic with respect to the Hadamard product, then \(G_1=G_2\). The following statement (theorem 3.9) is also very interesting: Suppose that there exists a domain \(\widetilde{G}\subset{\mathbb C}\) containing \(0\) such that \(H(\widetilde{G})\) is isomorphic to \(M(H(G))\). Then \(\widetilde{G}=\widehat{G}\) and the canonical injection \(L:H(\widehat{G})\to M(H(G))\) is already an isomorphism. The article contains three sections. In the first one the authors give equivalent operator-theoretic characterizations for the coefficient multipliers. In the second section the authors show that \(M(H(G))\) posesses a so-called strongly orthogonal sequence. It follows that an isomorphism on \(M(H(G))\) permutes the Taylor coefficients of the power series. The main results mentioned above are contained in the third section.











This page was built for publication: Multipliers on vector spaces of holomorphic functions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4519050)