English

Power means of probability measures and Ando-Hiai inequality

Functional Analysis 2019-09-24 v2

Abstract

Let μ\mu be a probability measure of compact support on the set Pn\mathbb{P}_n of all positive definite matrices, let t(0,1]t\in(0,1], and let Pt(μ)P_t(\mu) be the unique positive solution of X=PnXtZdμ(Z)X=\int_{\mathbb{P}_n}X\sharp_t Z d\mu(Z). In this paper, we show that Pt(μ)IPtp(ν)Pt(μ) P_t(\mu)\leq I\quad \Longrightarrow\quad P_{\frac{t}{p}}(\nu)\leq P_t(\mu) for every p1p\geq1, where ν(Z)=μ(Z1/p)\nu(Z)=\mu(Z^{1/p}). This provides an extension of the Ando--Hiai inequality for matrix power means. Moreover, we prove that if Φ:MnMm\Phi:\mathbb{M}_n\to\mathbb{M}_m is a unital positive linear map, then Φ(Pt(μ))Pt(ν)\Phi(P_t(\mu))\leq P_t(\nu) for all t[1,1]\{0}t\in[-1,1]\backslash\{0\}, where ν\nu is a certain measure.

Keywords

Cite

@article{arxiv.1806.04210,
  title  = {Power means of probability measures and Ando-Hiai inequality},
  author = {Mohsen Kian and Mohammad Sal Moslehian},
  journal= {arXiv preprint arXiv:1806.04210},
  year   = {2019}
}

Comments

Submitted on 24 November 2018