De Branges' conjecture on bounded Riemann mapping coefficients (Q913010): Difference between revisions

The Carathéodory-Fejér problem consists of finding necessary and sufficient conditions for the complex numbers \(a_ 0,a_ 1,...a_{k- 1}\) to be the first k coefficients of a power series that represents a function bounded by 1 in the open unit disk. L. de Branges formulated this problem in terms of r-equivalence of power series: Two power series \(f(z)=\sum a_ nz^ n\) and \(g(z)=\sum b_ nz^ n\) are r-equivalent if \(a_ n=b_ n\) for \(n=0,1,...,r-1\). He introduced the Hilbert space of r-equivalence classes of power series with norm \(\| f\|^ 2=\sum^{r-1}_{n=0}| a_ n|^ 2\), and he showed that a power series is r-equivalent to a power series that is bounded by 1 in the unit disk D if and only if the multiplication operator is a contraction on this Hilbert space. De Branges conjectured that a power series \(f(z)=\sum^{\infty}_{n=1}a_ nz^ n\) \((a_ 1>0)\) is r-equivalent to the power series of a normalized Riemann mapping function, which maps D onto a subregion of D, if the composition operator is a contraction on the space of r-equivalence classes of generalized power series and if some other conditions are satisfied. The converse of this conjecture is known to be true. The author disproves de Branges' conjecture for the case \(r=2\). This is done by computing various coefficients and examining them carefully.