Semi-discrete linear hyperbolic polyharmonic flows of closed polygons
James McCoy, Jahne Meyer
TL;DR
This work defines and analyzes a semi-discrete, second-order in time linear hyperbolic flow for polygons: $\frac{d^2\vec{X}}{dt^2} + \beta \frac{d\vec{X}}{dt} = (-1)^{m+1} M^m\vec{X}$, with $M$ circulant and amenable to Fourier diagonalization. It proves that arbitrary polygons can be evolved to a prescribed target polygon (Yau-type problem) and, after suitable rescaling, converge to an affine image of a regular $n$-gon; with $\beta>0$ the flow decays exponentially to a point, with the asymptotic shape governed by dominant eigenmodes with eigenvalues $\lambda_{m,k} = -4^m[\sin^2(\pi k/n)]^m$. The paper provides explicit planar and higher-codimension self-similar solutions via the $P_k$ modes, and extends the analysis to higher codimension polygons, showing analogous convergence behavior. It also introduces a semi-discrete Yau-type difference flow between polygons, proving long-time existence and convergence to the target, and discusses handling mismatched vertex counts and moving targets, illustrating the results with representative cases. These results advance understanding of linear hyperbolic flows in discrete polygon geometry and suggest new discrete curvature-flow analogies with potential applications.
Abstract
We consider the damped hyperbolic motion of polygons by a linear semi-discrete analogue of polyharmonic curve diffusion. We show that such flows may transition any polygon to any other polygon, reminiscent of the Yau problem of evolving one curve to another by a curvature flow, before converging exponentially to a point that, under appropriate rescaling, is a planar basis polygon. We also consider a hyperbolic linear semi-discrete flow of the Yau curvature difference flow, where a polygonal curve is able to flow to any other such that we get convergence to the target polygon in infinite time.
