Dobrushin’s Ergodicity Coefficient for Markov Operators on Cones
- First Online:
- 116 Downloads
Doeblin and Dobrushin characterized the contraction rate of Markov operators with respect the total variation norm. We generalize their results by giving an explicit formula for the contraction rate of a Markov operator over a cone in terms of pairs of extreme points with disjoint support in a set of abstract probability measures. By duality, we derive a characterization of the contraction rate of consensus dynamics over a cone with respect to Hopf’s oscillation seminorm (the infinitesimal seminorm associated with Hilbert’s projective metric). We apply these results to Kraus maps (noncommutative Markov chains, representing quantum channels), and characterize the ultimate contraction of the map in terms of the existence of a rank one matrix in a certain subspace.
KeywordsMarkov operator Dobrushin’s ergodicity coefficient ordered linear space invariant measure contraction ratio consensus noncommutative Markov chain quantum channel zero error capacity rank one matrix
Unable to display preview. Download preview PDF.