Rigidly breaking potential flows and a countable Alexandrov theorem for polytopes

TL;DR

This paper characterizes all possible rigid breaking motions of convex bodies in multiple dimensions using convex potentials, classifies these potentials via a countable Alexandrov theorem, and explores their connection to optimal transport and fractal examples.

Contribution

It introduces a classification of convex potentials governing rigid body breakings and extends Alexandrov's theorem to countably many pieces, linking to optimal transport theory.

Findings

Classification of potentials for rigid body breakings.

Establishment of a countable Alexandrov theorem for polytopes.

Connection between velocity fields and Wasserstein geodesics.

Abstract

We study all the ways that a given convex body in $d$ dimensions can break into countably many pieces that move away from each other rigidly at constant velocity, with no rotation or shearing. The initial velocity field is locally constant, but may be continuous and/or fail to be integrable. For any choice of mass-velocity pairs for the pieces, such a motion can be generated by the gradient of a convex potential that is affine on each piece. We classify such potentials in terms of a countable version of a theorem of Alexandrov for convex polytopes, and prove a stability theorem. For bounded velocities, there is a bijection between the mass-velocity data and optimal transport flows (Wasserstein geodesics) that are locally incompressible. Given any rigidly breaking velocity field that is the gradient of a continuous potential, the convexity of the potential is established under any of several conditions, such as the velocity field being continuous, the potential being semi-convex, the mass measure generated by a convexified transport potential being absolutely continuous, or there being a finite number of pieces. Also we describe a number of curious and paradoxical examples having fractal structure.