Extreme points and support points of subordination families (Q1589954)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Extreme points and support points of subordination families |
scientific article |
Statements
Extreme points and support points of subordination families (English)
0 references
28 May 2001
0 references
The family \(A\) of analytic functions \(f\) of the unit disk \(\Delta\), equipped with the topology of uniform convergence in compact subsets forms a locally convex linear space. Let \(B_0:=\{\omega\in A |\omega(0)=0, |\omega|<1\}\). A function \(f\in A\) is called \textit{subordinate} to \(g\), denoted by \(f\prec g\), if there exists a function \(\omega\in B_0\) such that \(f=g\circ \omega\). If \(g\) is univalent, this is equivalent to the normalization \(f(0)=g(0)\) and the geometric property \(f(\Delta)\subset g(\Delta)\). The family \(s(F):=\{f\in A |f\prec F\}\) is called a subordination family. The given article surveys known results for subordination families. Special cases are \(B_0=s(z)\) and \[ P:=\{p\in A |\text{ Re} p(z)>0\}=\{p\in A |p\prec {1+z \over 1-z}\} =s({1+z \over 1-z}) . \] The Krein-Milman theorem guarantees that a subordination family has \textit{extreme points} \(\text{ E} s(F)\neq \emptyset\), and these reconstruct \(s(F)\), i.~e. \(\overline{\text{ co}} \text{ E} s(F)=\overline{\text{ co}} s(F)\), \(\overline{\text{ co}} M\) denoting the closed convex hull of \(M\). The Herglotz representation formula shows via Choquet's theorem that \[ \text{ E} P=\{{1+xz \over 1-xz} ||x|=1\}. \] It is easy to see that in subordination families for every extreme point \(f\in \text{ E} s(F)\) (or \(f\in \text{ E} \overline{\text{ co}} s(F)\)), and \(f\prec g\), it follows that also \(g\in \text{ E} s(F)\) (or \(g\in \text{ E} \overline{\text{ co}} s(F)\), respectively) [W. Koepf: Extrempunkte und Stützpunkte in Familien nicht-verschwindender analytischer Funktionen; \(PhD\) thesis, Free University Berlin (1984)]. Hence the extreme points are large with respect to \(\prec\). A similar statement holds for the \textit{support points} of a linear functional \(L\). In the case of \(P\), the set of extreme points is smallest possible and consists only of the maximal elements w.~r.~t.\ \(\prec\). The authors show under which conditions for \(F\) this situation occurs. In contrast, for \(B_0\) one has \[ \text{ E} B_0= \{f\in B_0 |\int_0^{2\pi} \log (1-|f(e^{it})|) dt=-\infty\} \] Hence in this case, the set of extreme points is rather large. The article informs about the situation for general \(F\) which lies between the two extremes. Furthermore the set of support points of subordination families is considered. Finally, extreme point results for functions subordinate to convex, starlike and close-to-convex functions are given. Similar results for functions normalized by \(f(0)=1\) and \(f\neq 0\) can be found in [Complex Variables, Theory Appl. 8, 153-171 (1987; Zbl 0622.30002)].
0 references
Krein-Milman theorem
0 references
Herglotz representation formula
0 references
Choquet theorem
0 references
closed convex hull
0 references
univalent functions
0 references
convex functions
0 references
starlike functions
0 references
close-to-convex functions
0 references
0 references
0 references