Hereditarily frequently hypercyclic operators and disjoint frequent hypercyclicity

TL;DR

This paper introduces hereditary frequent hypercyclicity, explores its properties, and examines disjoint hypercyclicity, providing new results on operator classes, direct sums, and specific examples in functional analysis.

Contribution

It defines hereditary frequent hypercyclicity, establishes its properties, and applies it to study direct sums and disjoint hypercyclicity of operators, including new examples and counterexamples.

Findings

Operators satisfying the Frequent Hypercyclicity Criterion are hereditarily frequently hypercyclic.

Constructed examples of hypercyclic operators whose direct sums are not hypercyclic.

Extended disjoint hypercyclicity results for N-tuples of operators.

Abstract

We introduce and study the notion of hereditary frequent hypercyclicity, which is a reinforcement of the well known concept of frequent hypercyclicity. This notion is useful for the study of the dynamical properties of direct sums of operators; in particular, a basic observation is that the direct sum of a hereditarily frequently hypercyclic operator with any frequently hypercyclic operator is frequently hypercyclic. Among other results, we show that operators satisfying the Frequent Hypercyclicity Criterion are hereditarily frequently hypercyclic, as well as a large class of operators whose unimodular eigenvectors are spanning with respect to the Lebesgue measure. On the other hand, we exhibit two frequently hypercyclic weighted shifts $B_w,B_{w'}$ on $c_0(\mathbb{Z}_+)$ whose direct sum $B_w\oplus B_{w'}$ is not $\mathcal{U}$-frequently hypercyclic (so that neither of them is hereditarily frequently hypercyclic), and we construct a $C$-type operator on $\ell_p(\mathbb{Z}_+)$, $1\le p<\infty$ which is frequently hypercyclic but not hereditarily frequently hypercyclic. We also solve several problems concerning disjoint frequent hypercyclicity: we show that for every $N\in\mathbb{N}$, any disjoint frequently hypercyclic $N$-tuple of operators $(T_1,\dots ,T_N)$ can be extended to a disjoint frequently hypercyclic $(N+1)$-tuple $(T_1,\dots ,T_N, T_{N+1})$ as soon as the underlying space supports a hereditarily frequently hypercyclic operator; we construct a disjoint frequently hypercyclic pair which is not densely disjoint hypercyclic; and we show that the pair $(D,\tau_a)$ is disjoint frequently hypercyclic, where $D$ is the derivation operator acting on the space of entire functions and $\tau_a$ is the operator of translation by $a\in\mathbb{C}\setminus\{ 0\}$. Part of our results are in fact obtained in the general setting of Furstenberg families.