A hypercyclic operator whose adjoint is also hypercyclic
Héctor Salas
- 01 Mar 1991
- Vol. 112, Iss: 3, pp 765-770
TL;DR: In this paper, the existence of a bilateral weighted shift whose adjoint is also hypercyclic was shown, based on the similarity orbit theorem, which is a key result of Herrero's work.
read more
Abstract: An operator T acting on a Hubert space H is hypercyclic if, for some vector x in H, the orbit {Tnx: n > 0} is dense in H. We show the existence of a hypercyclic operator—in fact, a bilateral weighted shift—whose adjoint is also hypercyclic. This answers positively a question of D. A. Herrero. Introduction Let H be a complex, separable, infinite-dimensional Hubert space. Let B(H) denote the bounded linear operators acting on H. Let T e B(H) and x e H. The orbit of x under T is Orb(7\ x) — {x, Tx ,T x, ...} . T is said to be cyclic if there exists y e H such that the linear span of Oxb(T, y) is dense in H ; T is hypercyclic if Orb(T, y) itself is dense in H. In this last case, y is hypercyclic for T (in [2] such a y is called universal, and in [4] such a y is called orbital). In a recent paper [7], D. A. Herrero completely characterized the closure of the class of hypercyclic operators in B(H) in terms of spectral properties. His proof is based on previous work of C. Kitai [10], Godefroy and Shapiro [3], and Herrero [5] and [6]. Also a key result that he uses is the similarity orbit theorem [1]. In [7], Herrero raises several questions. The purpose of this note is to answer positively Question 2: "Does there exist T e B(H) such that both T and T* are hypercyclic?" We also answer in the affirmative Question 3: "Does there exist a hypercyclic backward shift such that bQbx ■ ■ ■ bk_x does not tend to infinity? (Here the b's are the weights of the shift.)" Gethner and Shapiro proved in [2] that if the weights bk > 1 and b0--bk go to infinity, then the backward shift is hypercyclic. In [7], it is shown that the condition bk > 1 can be dropped. Observe that if U is the unilateral (unweighted) shift, 2U* is hypercyclic, while 2I7 is not. Received by the editors January 2, 1990 and, in revised form, April 18, 1990. 1980 Mathematics Subject Classification (1985 Revision). Primary 47B99; Secondary 47B37.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
Hypercyclic weighted shifts
TL;DR: In this article, Chan et al. showed that I + T is hypercyclic whenever T is a unilateral backward weighted shift, thus answering in more generality a question recently posed by K C. Chan and J. H. Shapiro.
Hereditarily Hypercyclic Operators
Juan Bès,Alfredo Peris +1 more
TL;DR: In this article, it was shown that a continuous linear operator T on a Frechet space satisfies the Hypercyclicity Criterion if and only if it is hereditarily hypercyclic.
306
Limits of hypercyclic and supercyclic operators
TL;DR: In this article, it was shown that if an operator T ∗ does not have any eigenvalue and that the union of the spectrum of T and the unit circle is a connected set, then it is hypercyclic.
205
Recent developments in hypercyclicity
Karl-Goswin Grosse-Erdmann
- 01 Jan 2003
TL;DR: In these notes, recent progress in the theory of hypercyclic and chaotic operators is reported on, guided by the following fundamental problems: How do the authors recognize hyper cyclic operators?
References
Limits of hypercyclic and supercyclic operators
TL;DR: In this article, it was shown that if an operator T ∗ does not have any eigenvalue and that the union of the spectrum of T and the unit circle is a connected set, then it is hypercyclic.
205
The diagonal entries in the formula “quasitriangular - compact = triangular” and restrictions of quasitriangularity
Abstract: . A (bounded linear) Hilbert space operator T is called quasi triangular if there exists an increasing sequence {P n }::'-o of finite-rank orthogonal projections, converging strongly to 1, such that 11<1 - Pn)TPnll ..... 0 (n ..... 00). This definition, due to P. R. Halmos, plays a very important role in operator theory. The core of this article is a concrete answer to the following problem: Suppose T is a quasi triangular operator and r = {A j }j'-l is a sequence of complex numbers. Find necessary and sufficient conditions for the existence of a compact operator K (of arbitrarily small norm) so that T - K is triangular with respect to some orthonormal basis, and the sequence of diagonal entries of T - K coincides with r. For instance, if no restrictions are put on the norm of K, then T and r must be related as follows: (a) if Ao is a limit point of r and Ao - T is semi-Fredholm, then ind(Ao - T) > 0; and (b) if n is an open set intersecting the Weyl spectrum of T, whose boundary does not intersect this set, then {j: A j En} is a denumerable set of indices. Particularly important is the case when r = {O, 0, 0, ... }. The following are equivalent for an operator T: (1) there is an integral sequence {P n }::'=o of orthogonal projections, with rank p. = n for all n, converging strongly to 1, such that 11<1 - Pn)TP.+11I ..... 0 (n ..... 00); (2) from some compact K, T - K is triangular, with diagonal entries equal to 0; (3) Tis quasitriangular, and the Weyl spectrum of T is connected and contains the origin. The family (StrQT)_l of all operators