In-betweenness, a geometrical monotonicity property for operator means (Q1940326): Difference between revisions

From MaRDI portal
Set OpenAlex properties.
ReferenceBot (talk | contribs)
Changed an Item
 
(One intermediate revision by one other user not shown)
Property / arXiv ID
 
Property / arXiv ID: 1011.6313 / rank
 
Normal rank
Property / cites work
 
Property / cites work: Series and parallel addition of matrices / rank
 
Normal rank
Property / cites work
 
Property / cites work: Asymptotic error rates in quantum hypothesis testing / rank
 
Normal rank
Property / cites work
 
Property / cites work: Inequalities between means of positive operators / rank
 
Normal rank
Property / cites work
 
Property / cites work: Analysis 2 / rank
 
Normal rank
Property / cites work
 
Property / cites work: On the exponential metric increasing property. / rank
 
Normal rank
Property / cites work
 
Property / cites work: Interpolating the arithmetic--geometric mean inequality and its operator version / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q4241835 / rank
 
Normal rank
Property / cites work
 
Property / cites work: Means and concave products of positive semi-definite matrices / rank
 
Normal rank
Property / cites work
 
Property / cites work: Means of positive linear operators / rank
 
Normal rank

Latest revision as of 06:46, 6 July 2024

scientific article
Language Label Description Also known as
English
In-betweenness, a geometrical monotonicity property for operator means
scientific article

    Statements

    In-betweenness, a geometrical monotonicity property for operator means (English)
    0 references
    6 March 2013
    0 references
    The in-betweenness property of a scalar mean \(\mu\) means that if \(x \leq y\), then \(x \leq \mu(x,y)\leq y\). The author introduces the notion of in-betweenness with respect to a metric for operator means, overcoming the lack of a total ordering and of the existence of a natural metric. He defines in-betweenness w.r.t. Euclidean distance and in-betweenness w.r.t. angle, which are based on endowing the set of trace class operators with the Euclidean metric induced by the Hilbert-Schmidt inner product \(\langle A, B\rangle = \text{Tr}[A^*B]\). The author calls a weighted operator mean \(\mu(A, B, t)\) distance-monotonic (angle-monotonic, resp.) if the Euclidean distance (the angle, resp.) between \(A\) and \(\mu(A, B, t)\) decreases monotonically with \(t \in [0, 1]\). He then investigates the monotonicity of the power means \(((A^p+B^p)/2)^{1/p}\,\,(p \in \mathbb{R})\) and the Heinz means \(A^\nu B^{1-\nu}\,\,(1 \leq \nu\leq 1)\). He further exhibits a counterexample, showing that Kubo-Ando means (cf. \textit{F. Kubo} and \textit{T. Ando} [Math. Ann. 246, 205--224 (1980; Zbl 0412.47013)]) generally do not satisfy in-betweenness with respect to the Euclidean distance. In contrast, the author proves that they all satisfy in-betweenness with respect to the trace metric distance \(\delta(A, B) = \|\log(A^{-1/2}BA^{-1/2})\|^2\).
    0 references
    0 references
    0 references
    0 references
    0 references
    power means
    0 references
    Heinz means
    0 references
    Kubo-Ando means
    0 references
    monotonicity
    0 references
    in-betweenness
    0 references
    angle
    0 references
    trace
    0 references
    0 references
    0 references