Julia sets of random exponential maps

TL;DR

This paper investigates the Julia sets of non-autonomous exponential maps with varying parameters, showing that under certain conditions, the Julia set is the entire complex plane, and exploring how it depends on the sequence of parameters.

Contribution

It introduces the study of Julia sets for non-autonomous exponential maps and proves results about their size and properties for various parameter sequences.

Findings

Julia set is the whole plane for random sequences near 1/e with probability 1

Julia set is the whole plane for sequences of the form 1/e + 1/n^p with p<1/2

Existence of sequences where orbits tend to infinity but the Fatou set is non-empty

Abstract

For a sequence $(\lambda_n)$ of positive real numbers we consider the exponential functions $f_{\lambda_n} (z) = \lambda_n e^z$ and the compositions $F_n = f_{\lambda_n} \circ f_{\lambda_{n-1}} \circ ... \circ f_{\lambda_1}$. For such a non-autonomous family we can define the Fatou and Julia sets analogously to the usual case of autonomous iteration. The aim of this document is to study how the Julia set depends on the sequence $(\lambda_n)$. Among other results, we prove the Julia set for a random sequence $\{\lambda_n \}$, chosen uniformly from a neighbourhood of $\frac{1}{e}$, is the whole plane with probability $1$. We also prove the Julia set for $\frac{1}{e} + \frac{1}{n^p}$ is the whole plane for $p < \frac{1}{2}$, and give an example of a sequence $\{\lambda_n \} $ for which the iterates of $0$ converge to infinity starting from any index, but the Fatou set is non-empty.