New dimensional bounds for a branched transport problem
Alessandro Cosenza, Michael Goldman, Melanie Koser
TL;DR
The paper analyzes a reduced branched transport model with a Sobolev boundary penalty to capture boundary patterning in type-I superconductors. It develops a framework linking local energy scaling to fractal dimensions of the irrigated boundary measure, leveraging McCann interpolation, the no-loop property in 1D, and Ahlfors regularity assumptions. The main results show that for s \\leq 1/4 the boundary measure has lower dimension 1, while for s \\gt 1/4 the dimension is pinned by ᾱ depending on regularity: dim_{F} \\mu_T \\leq ᾱ and dim_M \\mu_T \\geq ᾱ, with sharp conclusions when mu_T is upper or lower \\alpha-Ahlfors regular; in particular, mu_T cannot be smooth. By combining local energy bounds with a first variation argument, the authors prove a rigorous fractal-type lower bound and identify regimes where the boundary exhibits non-integer dimensionality, offering a rigorous justification for fractal boundary behavior in this branched transport reduction.
Abstract
We consider a branched transport problem with weakly imposed boundary conditions. This problem arises as a reduced model for pattern formation in type-I superconductors. For this model, it is conjectured that the dimension of the boundary measure is non-integer. We prove this conjecture in a simplified 2D setting, under the (strong) assumption of Ahlfors regularity of the irrigated measure. This work is the first rigorous proof of a singular behaviour for irrigated measures resulting from minimality.
