On the invariance of super-linearization under polynomial automorphisms
Anmol Harshana, Mohamed-Ali Belabbas
TL;DR
Addresses whether the class $\mathcal{S}_n$ of super-linearizable polynomial vector fields is preserved under polynomial automorphisms. Develops a proof framework showing invariance under tame automorphisms (affine and elementary steps) and analyzes the weighted dependency graph (WDG) condition as a sufficient criterion for super-linearization, noting its noninvariance under some automorphisms. Shows that the WDG property is not generally preserved under elementary transformations, but lifting with a stabilizing observable can restore WDG compliance after a coordinate change. The results clarify how coordinate changes interact with lifted representations and guide the use of global linearization approaches in dynamical analysis, with implications for Koopman-based methods and polyflow representations.
Abstract
We prove that the super-linearizability of polynomial systems is preserved by all currently known classes of polynomial automorphisms of $\R^n$. We then establish connections between such automorphisms and a sufficient condition for super-linearizability.
