Convex univalent functions and continuous linear right inverses

Topological linear spaces of continuous, differentiable or analytic functions (46E10) General (adjoints, conjugates, products, inverses, domains, ranges, etc.) (47A05)

Review text

If \(G\) is a convex open subset of the complex plane \(C\) and if \(F(z)=\sum_{j=0} ^ \infty a_ j x^ j\) is a non-zero entire function of finite type and at most order one, then the mapping \[ L_ F(f)=\sum_{j=0}^ \infty a_ j f^{(j)},\text{ for \(f\) analytic on } G, \] defines a surjective continuous linear differential operator from the space \(A(G)\) of all functions analytic on \(G\) onto itself. The general problem considered in the paper under review is to determine conditions under which each such \(L_ F\) will have a continuous linear right inverse; that is, for each such \(F\) one wishes to find a function \(R\) on \(A(G)\) satisfying \(L_ F(R(g))=g\) for all \(g\) and such that \(R\) depends linearly and continuously on \(g\). Earlier works of Schwerdtfeger, Taylor, and Momm are cited which showed that in case \(G=\mathbb{C}\) or \(G\) is an open disk then each such \(L_ F\) has the desired right inverse. The current paper addresses itself to the case of an arbitrary open convex set \(G\neq \mathbb{C}\). An important role is played by the analytic and univalent function \(\psi\) mapping the open unit disc onto \(G\) and specifically by the support function \(h_ r\), of the set \(\{\psi(z)\); \(| z|<r\}\); namely, for \(t\) real, one has \(h_ r(t)=\sup\{\text{Re}(\psi(z) e^{-it})\): \(| z|<r\}\). Two principal results are presented. Theorem 3.1 asserts that, taking \(G\neq\mathbb{C}\) to be an open convex set and \(F\) a non-zero entire function of order at most one and minimal type, then the surjective differential operator \(L_ F\) on \(A(G)\) admits a continuous linear right inverse if, and only if, there is a number \(C>0\) with \[ d(-t)=\lim_{r\to 1} {{h_ 1(-t)-h_ r(-t)} \over {1-r}}\leq C \] for all accumulation points \(t\) of the set \(\{\arg(a)\): \(F(a)=0\}\). This leads through Lemma 3.4 (where much of the work is contained) to Theorem 3.5 which establishes the equivalence of the following four statements (\(G\) and \(F\) are as in Theorem 3.1). (i) Each non-zero differential operator \(L_ F\) on \(A(G)\) has a continuous linear right inverse. (ii) The function \(d(s)=\lim_{r\to 1}(h_ 1(s)-h_ r(s))/(1-r)\) is bounded on the reals. (iii) The function \(\psi'\) is bounded. (iv) The set \(G\) is bounded and the angular derivatives of \(\psi\) at points on the boundary of the unit disc are uniformly bounded. The paper concludes with a discussion of some examples and a demonstration that, for a given \(G\), functions \(F\) for which \(L_ F\) has a right inverse always exist.

Recommendations

Cites work

Cited in

(16)

This page was built for publication: Convex univalent functions and continuous linear right inverses

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1185745)