Zeros of sections of power series: deterministic and random (Q1687103): Difference between revisions

The paper deals with a class of power series whose radius of convergence is \(1\). To each of its non-constant sections \(s_n=s_n(z)=\sum_{k=0}^{n} a_k z^k, \;n=0, 1,2,\ldots\), two measures are associated: the zero counting measure \(\mu_n(f)\) and its circular projection \(\rho_n(f)\), i.e. \[ \mu_n=\frac{1}{n} \sum_{\omega:\;s_n(\omega)=0} \delta_{\omega} \quad \text{and} \quad \rho_n=\rho_n(f)=\frac{1}{n} \sum_{\omega:\;s_n(\omega)=0} \delta_{|\omega|}. \] The author presents a streamlined proof and some refinements of classical results on asymptotics of zero-counting measures, from the seminal work of R. Jentzsch and G. Szegő up to the characterization (due to F. Carlson and G. Bourion, and also P. Erdős and H. Fried) of the Szegő class \(\mathcal{S}\) of those holomorphic functions \(f\) whose zero counting measures weakly converge to the uniform probability \(\Lambda\). Then this characterization is applied to obtain some results (both known and new) on the asymptotic distribution of zeros of sections of random power series. The author analyzes the probability that a random power series is a Szegő power series. He also describes all the possible weak limit points of the probabilities \(\rho_n\) associated to a power series \(f\), all whose coefficients are either \(0\) or \(1\). In the last section of the paper the author exhibits an example of a universal power series \(f\) such that every probability measure in \([1,\;+\infty)\) is a limit of a subsequence of the probability measures \((\rho_n(f))_{n\geq 0}\).