Phase transitions in isoperimetric problems on the integers
Joseph Briggs, Chris Wells
TL;DR
This work shows that the natural nesting of extremal sets for vertex- and edge-isoperimetry on Cayley graphs of $\mathbb{Z}$ fails already in dimension one, via a phase-transition mechanism that mimics cylindrical grids. It constructs non-nested extremal families by relating $\mathbb{Z}$-Cayley graphs to cylinder-like embeddings and proves that intervals become eventually optimal for large sizes on any Cayley graph $G=Cay(\mathbb{Z},B)$ generating $\mathbb{Z}$, thereby answering Barber–Erde’s Question 2 positively in dimension one when finitely many sets are ignored. The paper provides precise thresholds: for edge- and vertex-problems on $G= Cay(\mathbb{Z},\pm\{1,b\})$ and $Cay(\mathbb{Z},\pm\{1,b-1,b,b+1\})$, optimizers are intervals for large $n$, but can be translates of $[k]+b[k]$ for certain square sizes $n=k^2$ when $k<(b+1)/2$ or $(b-1)/2$ respectively. It also establishes that intervals are eventually optimal for any Cayley graph on $\mathbb{Z}$, using residue-class and compression arguments, and discusses implications and open problems for higher dimensions.
Abstract
Barber and Erde asked the following question: if $B$ generates $\mathbb Z^d$ as an additive group, then must the extremal sets for the vertex/edge-isoperimetric inequality on the Cayley graph $\operatorname{Cay}(\mathbb Z^d,B)$ form a nested family? We answer this question negatively for both the vertex- and edge-isoperimetric inequalities, specifically in the case of $d=1$. The key is to show that the structure of the cylinder $\mathbb Z\times(\mathbb Z/k\mathbb Z)$ can be mimicked in certain Cayley graphs on $\mathbb Z$, leading to a phase transition. We do, however, show that Barber--Erde's question for Cayley graphs on $\mathbb Z$ has a positive answer if one is allowed to ignore finitely many sets.
