The iterated Aluthge transforms of a matrix converge (Q616894)

From MaRDI portal
scientific article
Language Label Description Also known as
English
The iterated Aluthge transforms of a matrix converge
scientific article

    Statements

    The iterated Aluthge transforms of a matrix converge (English)
    0 references
    0 references
    0 references
    0 references
    12 January 2011
    0 references
    Let \(\mathcal{H}\) be a Hilbert space and \(T\) a bounded operator defined on \(\mathcal{H}\) whose polar decomposition is \(T=U|T|\). The Aluthge transform of \(T\) is the operator \(\Delta(T)=|T|^{1/2}U|T|^{1/2}\). This transform was introduced in [\textit{A. Aluthge}, Integral Equations Oper. Theory 13, No.~3, 307--315 (1990; Zbl 0718.47015)] to study \(p\)-hyponormal and log-hyponormal operators. Roughly speaking, the idea behind the Aluthge transform is to take an operator into another operator which shares some spectral properties with the first one, in particular the spectrum, and it is closer to be normal. It is also well known that \(\Delta(\lambda VTV^*)=\lambda V\Delta(T)V^*\) for every \(\lambda \in \mathbb{C}\) and unitary operator \(V\), that it is a contraction (not necessary strict) with respect to the spectral norm, and that \[ \Delta(T_1 \oplus T_2)= \Delta(T_1) \oplus \Delta(T_2). \tag{1} \] The Aluthge transform has received much attention recently. One reason is its connection with the invariant subspace problem. \textit{I. B. Jung, E. Ko} and \textit{C. Pearcy} proved in [Integral Equations Oper.\ Theory 37, No.~4, 437--448 (2000; Zbl 0996.47008)] that \(T\) has a nontrivial invariant subspace if and only if \(\Delta (T)\) does. On the other hand, \textit{Dykema} and \textit{Schultz} [On Aluthge transform: continuity properties and Brown measure, preprint available in \url{arXiv:math/0512197}] proved that the Brown measure is preserved by the Aluthge transform. Another reason is related with the iterated Aluthge transform. Let \(\Delta^n(T)\) denote the \(n\)-times iterated Aluthge transform of \(T\), i.e., \(\Delta^0(T)=T\) and \(\Delta^n(T)=\Delta(\Delta^{n-1}(T))\) for every \(n\in \mathbb{N}\). In [Integral Equations Oper. Theory 45, No.~4, 375--387 (2003; Zbl 1030.47021)] \textit{I. B. Jung, E. Ko} and \textit{C. Pearcy} raised the following conjecture: Conjecture 1. For every bounded operator \(T\) on \(\mathcal{H}\), the sequence of iterates \(\{\Delta^n(T)\}_{n\in \mathbb{N}}\) convergence. The main goal of this article is to prove that the Conjecture 1. is true for every matrix, and hence to completely solve the problem in the finite dimensional case. The suggested bination of dynamical system techniques with geometrical arguments. The dynamical system techniques used here have already shown to be useful for the study of this problem. In the previous work [\textit{J. Antezana, E. R. Puljals} and \textit{D. Stojanoff}, Adv. Math. 216, No.~1, 255--278 (2007; Zbl 1129.37009)] the authors were used to prove the convergence of the iterated Aluthge transform sequence for diagonalizable matrices.
    0 references
    0 references
    Aluthge transform
    0 references
    stable manifold theorem
    0 references
    similarity orbit
    0 references
    polar decomposition
    0 references
    0 references
    0 references
    0 references