An application of Brouwer's fixed-point theorem: continuously differentiable convex functions with gradient of constant norm

TL;DR

This paper proves that a continuously differentiable convex function with a constant gradient norm must be affine, providing a first-order characterization without second-order conditions, using Brouwer's fixed-point theorem.

Contribution

It introduces a novel first-order characterization of affine functions among convex functions, based solely on gradient properties and Brouwer's theorem, without second-order assumptions.

Findings

Convex functions with constant gradient norm are affine

The proof relies on Brouwer's fixed-point theorem and inequalities

Potential generalization to Hilbert space functions

Abstract

As an application of Brouwer's fixed-point theorem we prove that a continuously differentiable convex function with gradient of constant norm is an affine mapping. It is a first-order characterization of affine mappings among continuously differentiable convex functions, because neither the second-order condition of convexity nor related operators are used. The condition of differentiability is essential as the case of the norm function shows. In addition to Brouwer's theorem, the proof is based on the Cauchy--Bunyakovsky--Schwarz inequality and becomes complete by minimizing the distance between lines of gradient directions. Following the steps of the proof, we sketch a possible generalization of the result to functions defined on Hilbert spaces.