Operator means of probability measures and generalized Karcher equations

DOI10.1016/J.AIM.2015.11.019MaRDI QIDQ904036zbMATH OpenOpenAlexFDO

Publication date 15 January 2016

Published in Advances in Mathematics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1601.06777

Karcher mean operator mean operator monotone function

General theory of (C^*)-algebras (46L05) Operator means involving linear operators, shorted linear operators, etc. (47A64) Global Riemannian geometry, including pinching (53C20) Differential geometry of symmetric spaces (53C35)

Abstract: In this article we consider means of positive bounded linear operators on a Hilbert space. We present a complete theory that provides a framework which extends the theory of the Karcher mean, its approximating matrix power means, and a large part of Kubo-Ando theory to arbitrary many variables, in fact, to the case of probability measures with bounded support on the cone of positive definite operators. This framework characterizes each operator mean extrinsically as unique solutions of generalized Karcher equations which are obtained by exchanging the matrix logarithm function in the Karcher equation to arbitrary operator monotone functions over the positive real half-line. If the underlying Hilbert space is finite dimensional, then these generalized Karcher equations are Riemannian gradients of convex combinations of strictly geodesically convex log-determinant divergence functions, hence these new means are the global minimizers of them, in analogue to the case of the Karcher mean as pointed out. Our framework is based on fundamental contraction results with respect to the Thompson metric, which provides us nonlinear contraction semigroups in the cone of positive definite operators that form a decreasing net approximating these operator means in the strong topology from above.

Review text

Let \(\mathbb P\) denote the cone of positive definite matrices of order \(n\) with complex entries. The \textit{weighted multivariable geometric mean} or \textit{Karcher mean} of the \(k\)-tuple \(\mathbb A:=(A_1,A_2,\dots,A_k)\in\mathbb P^k\) relative to the \textit{positive probability vector} \(\omega:=(w_1,w_2,\dots,w_k)\) is defined as the center of mass \(\Lambda(w_1,w_2,\dots,w_k;A_1,A_2,\dots, A_k):=\arg\min_X\sum_{i=1}^kw_id^2(X,A_i)\), where \(X\) runs over \(\mathbb P\). Here \(d\) denotes the Riemannian distance on the manifold \(\mathbb P\). It is known that the Karcher mean is the unique positive definite solution of the corresponding Karcher equation: \(\sum_{i=1}^kw_i \log(X^{-1}A_i)=0\). Among other things, the author extends the theory of the Karcher mean by taking any operator monotone function instead of the logarithmic function in the Karcher equation. Another generalization that is considered is to replace the sums by integrals with respect to probability measures supported on \(\mathbb P\). In order to prove his results, the author, among other things, deduces a generalized form of the Karcher mean via strictly geodesically convex divergence functions and obtains certain new integral formulae. The framework for these results is based on contraction results relative to the Thompson metric, which yield nonlinear contraction semigroups in the cone of positive definite operators.

Recommendations

Cites work

Cited in

(35)

This page was built for publication: Operator means of probability measures and generalized Karcher equations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q904036)