2
$\begingroup$

It is well-known that the weighted backshift operator $B_{\lambda}:\ell^p \rightarrow \ell^p$ is hypercyclic (with $\lambda>1$); that is, there exists a dense set of sequences $X\subseteq \ell^p$ for which $$ \overline{\left\{B^n(x)\right\}_{n \in \mathbb{N}} } = \ell^p \qquad (\forall x \in X). $$

Is there a known, concrete example of such an $x\in X$?

$\endgroup$
2
  • 2
    $\begingroup$ The backward shift itself is certainly not hypercyclic because it is a contaction. Multiples $\lambda B$ with $|\lambda|>1$ are hypercyclic. $\endgroup$ Commented Oct 2, 2019 at 13:43
  • $\begingroup$ I made the change, you're right. Thanks Jochen! $\endgroup$
    – ABIM
    Commented Oct 2, 2019 at 14:01

1 Answer 1

3
$\begingroup$

Probably not an example you are looking for. Rolewicz's proof is essentially an application of Baire's theorem and you can make this in a certain sense construtive: Choose a dense countable set $\{x^k: k\in \mathbb N\}$ of finite sequences (i.e., terminating with zeros) in $\ell^p$, align $y^k=(x^k_1,\lambda x^k_2, \lambda^2 x^k_3,\ldots)$ in a single sequence separated by many zeros and mutiply the $n$-th term of this sequence by $1/\lambda^{n-1}$. Choosing the blocks of separating zeros long enough will make this sequence an element of $\ell^p$. Given $k$ you apply $\lambda B$ many times to get the sequence $x^k$ in front followed by many zeros and a rest with very small $p$-norm.

The missing details are very ugly. The only advantage compared to the much more elegant use of Baire's theorem is that you can apply such a kind of construction to other locally convex sequence spaces as, e.g., in Bonet, J.; Frerick, L.; Peris, A.; Wengenroth, J. Transitive and hypercyclic operators on locally convex spaces, Bull. Lond. Math. Soc. 37, No. 2, 254-264 (2005).

$\endgroup$
5
  • $\begingroup$ Hi Jochen. I have seen a version of this proof in Karl Grosse-Erdmann's book. However, there is a point where it jumps from constructive to non-constructive and that's where it loses value for me. Thanks though I very much appreciate it :) $\endgroup$
    – ABIM
    Commented Oct 3, 2019 at 10:57
  • $\begingroup$ The authors of the book you mention are K.G. Grosse-Erdmann and A. Peris. $\endgroup$ Commented Oct 3, 2019 at 11:41
  • $\begingroup$ Precisely, both of them. $\endgroup$
    – ABIM
    Commented Oct 3, 2019 at 11:43
  • $\begingroup$ @AIM_BLB I think it's not constructive only if you make no effort to understand the construction. $\endgroup$
    – YCor
    Commented Oct 3, 2019 at 12:00
  • $\begingroup$ Ah, in that case I'll go through the details of the original paper and hopefully infer the description of the hypercylic vector $\endgroup$
    – ABIM
    Commented Oct 3, 2019 at 12:25

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .