On Bauschke-Bendit-Moursi modulus of averagedness and classifications of averaged nonexpansive operators

TL;DR

This paper introduces classifications of averaged and related operators using the Bauschke-Bendit-Moursi modulus, revealing new properties and explicit formulas for their analysis in convex optimization.

Contribution

It proposes novel classifications of averaged operators based on the Bauschke-Bendit-Moursi modulus and provides explicit formulas for their calculation.

Findings

Operators with averaging constant less than 1/2 are bi-Lipschitz homeomorphisms

Proximal operators of Lipschitz smooth functions have modulus less than 1/2

Explicit formulas for the modulus of averagedness of resolvents and proximal operators

Abstract

Averaged operators are important in Convex Analysis and Optimization Algorithms. In this paper, we propose classifications of averaged operators, firmly nonexpansive operators, and proximal operators using the Bauschke-Bendit-Moursi modulus of averagedness. We show that if an operator is averaged with a constant less than 1/2, then it is a bi-Lipschitz homeomorphism. Amazingly the proximal operator of a convex function has its modulus of averagedness less than 1/2 if and only if the function is Lipschitz smooth. Some results on the averagedness of operator compositions are obtained. Explicit formulae for calculating the modulus of averagedness of resolvents and proximal operators in terms of various values associated with the maximally monotone operator or subdifferential are also given. Examples are provided to illustrate our results.