Rigorous dense graph limit of a model for biological transportation networks
Nuno J. Alves, Jan Haskovec
TL;DR
This work rigorously derives the dense graph (graphon) limit of a discrete biological transport network model with pressure-driven flows and a convex energy combining pumping and metabolic costs under Kirchhoff mass conservation. By introducing a regularized, rescaled semi-discrete formulation and proving $\Gamma$-convergence to a continuum energy $\mathcal{F}$ defined on a graphon, the authors establish convergence of global minimizers from the discrete to the continuum problem via a nonlocal Poisson-type equation for the pressure. The analysis hinges on a careful treatment of the Kirchhoff law, the Poisson equation, and the kinetic/metabolic energy components, including convexity, lower semicontinuity, and strong-weak convergence arguments. The results provide a rigorous mathematical foundation for continuum descriptions of biological transport networks emerging from dense, discrete structures, clarifying the role of connectivity, graphon structure (0-1 valued), and scaling in the limiting process.
Abstract
We rigorously derive the dense graph limit of a discrete model describing the formation of biological transportation networks. The discrete model, defined on undirected graphs with pressure-driven flows, incorporates a convex energy functional combining pumping and metabolic costs. It is constrained by a Kirchhoff law reflecting the local mass conservation. We first rescale and reformulate the discrete energy functional as an integral `semi-discrete' functional, where the Kirchhoff law transforms into a nonlocal elliptic integral equation. Assuming that the sequence of graphs is uniformly connected and that the limiting graphon is 0-1 valued, we prove two results: (1) rigorous Gamma-convergence of the sequence of the semi-discrete functionals to a continuum limit as the number of graph nodes and edges tends to infinity; (2) convergence of global minimizers of the discrete functionals to a global minimizer of the limiting continuum functional. Our results provide a rigorous mathematical foundation for the continuum description of biological transport structures emerging from discrete networks.
