A sharp quantitative version of Hales' isoperimetric honeycomb theorem (Q311164)

From MaRDI portal
scientific article
Language Label Description Also known as
English
A sharp quantitative version of Hales' isoperimetric honeycomb theorem
scientific article

    Statements

    A sharp quantitative version of Hales' isoperimetric honeycomb theorem (English)
    0 references
    0 references
    0 references
    29 September 2016
    0 references
    A \(N\)-tiling \({\mathcal E}\) of a two-dimensional torus \({\mathcal T}\) is a family \({\mathcal E}=\{{\mathcal E}(h)\}^N_{h=1}\) of sets of finite perimeter in \({\mathcal T}\) such that \(|{\mathcal T}\setminus\bigcup^N_{h=1}{\mathcal E}(h)|=0\) and \(|{\mathcal E}(h)\cap{\mathcal E}(k)|=0\) for every \(h, k\in \mathbb N\), \(h\neq k\). The volume of \({\mathcal E}\) is \(\text{vol}\,{\mathcal E}=\{|{\mathcal E}(1)|,|{\mathcal E}(2)|,\dots,|{\mathcal E}(N)|\}\), and the relative perimeter of \({\mathcal E}\) in \(A\subset{\mathcal T}\) is given by \(P({\mathcal E},A)=\frac12\sum\limits_{h=1}^NP({\mathcal E}(h),A)\), where \(P(E,A)={\mathcal H}^1(A\cap\partial E)\) if \(E\) is an open set with Lipschitz boundary. The distance between two tilings \({\mathcal E}\) and \({\mathcal F}\) is defined as \(d({\mathcal E},{\mathcal F})=\frac12\sum\limits_{h=1}^N|{\mathcal E}(h)\Delta{\mathcal F}(h)|\). If \(H\) is a regular unit-area hexagon, then \({\mathcal H}=\{{\mathcal H}(h)\}^N_{h=1}\) is a reference unit-area tiling consisting of \(\alpha\) rows and \(\beta\) columns of regular hexagons by considering certain translations of \(H\). Hales' isoperimeiric honeycomb theorem states that \(P({\mathcal E})\geq P({\mathcal H})\) whenever \({\mathcal E}\) is a unit-area tiling of \({\mathcal T}\). In this paper, the authors prove a sharp quantitative version of Hales' isoperimetric honeycomb theorem by exploiting a quantitative isoperimetric inequality for polygons and an improved convergence theorem for planar bubble clusters. They show that there exists a positive constant \(\kappa\) depending on \({\mathcal T}\) such that \(P({\mathcal E})\geq P({\mathcal H})\left\{1+\kappa\alpha({\mathcal E})\right\}\) whenever \({\mathcal E}\) is a unit-area tiling of \({\mathcal T}\) and \(\alpha({\mathcal E})=\inf d(\widetilde{\mathcal E},v+{\mathcal H})\), where \(d\) is the distance between two tilings, and the minimization takes place among all \(v=(t\sqrt{3}\, l,s\,l)\), \(s,l\in[0,1]\), and among all tilings \(\widetilde{\mathcal E}\) obtained by setting \(\widetilde{\mathcal E}(h)={\mathcal E}(\sigma(h))\) for a permutation \(\sigma\) of \(\{1,\dots,N\}\). Also, the authors provide a description of isoperimetric tilings of the torus with respect to almost unit-area constraints or with respect to almost flat Riemannian metrics.
    0 references
    0 references
    0 references
    isoperimetric problems
    0 references
    partitioning problems
    0 references
    Hales' honeycomb theorem
    0 references
    0 references
    0 references