Dobrushin's ergodicity coefficient for Markov operators on cones (Q487212)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Dobrushin's ergodicity coefficient for Markov operators on cones
scientific article

    Statements

    Dobrushin's ergodicity coefficient for Markov operators on cones (English)
    0 references
    0 references
    0 references
    19 January 2015
    0 references
    The authors generalize results due to Doeblin and Dobrushin on the contraction rate of Markov operators with respect to the total variation norm. Such a generalization is given in terms of Markov operators over cones and their first main result is a theorem on the contraction rate in Hopf's oscillation seminorm. The authors discuss several motivations for considering cones which are different from the standard positive cone of \(\mathbb{R}^n\), with a relevant application being a noncommutative analogue of Dobrushin's inequality (a consequence of the main theorem) which is in turn applied to discuss ergodicity of noncommutative Markov chains. The latter is then characterized in terms of the existence of a rank one matrix in certain subspaces of matrices. Also results are proved in the setting of nonlinear maps over cones where, in particular, the authors are able to compute the contraction rate of a nonlinear map with respect to Hilbert's projective metric. This is accomplished by making use of one of their main theorems in combination with an expression obtained by \textit{R. D. Nussbaum} [Differ. Integral Equ. 7, No.~ 5--6, 1649--1707 (1994; Zbl 0844.58010)] for \(\kappa(x)\), a variant of the Doeblin-Dobrushin coefficient.
    0 references
    0 references
    Markov operator
    0 references
    Dobrushin's ergodicity coefficient
    0 references
    ordered linear space
    0 references
    invariant measure
    0 references
    contraction ratio
    0 references
    consensus
    0 references
    noncommutative Markov chain
    0 references
    quantum channel
    0 references
    rank one matrix
    0 references
    0 references
    0 references