The discrete Heisenberg group HZ2k+1 is the group generated by a1,b1,…,ak,bk,c, subject to the relations [a1,b1]=…=[ak,bk]=c and [ai,aj]=[bi,bj]=[ai,bj]=[ai,c]=[bi,c]=1 for every distinct i,j∈{1,…,k}. Denote S={a1±1,b1±1,…,ak±1,bk±1}. The horizontal boundary of Ω⊂HZ2k+1, denoted ∂hΩ, is the set of all (x,y)∈Ω×(HZ2k+1∖Ω) such that x−1y∈S. The horizontal perimeter of Ω is ∣∂hΩ∣. For t∈N, define ∂vtΩ to be the set of all (x,y)∈Ω×(HZ2k+1∖Ω) such that x−1y∈{ct,c−t}. The vertical perimeter of Ω is defined by ∣∂vΩ∣=∑t=1∞∣∂vtΩ∣2/t2. It is shown here that if k≥2, then ∣∂vΩ∣≲k1∣∂hΩ∣. The proof of this "vertical versus horizontal isoperimetric inequality" uses a new structural result that decomposes sets of finite perimeter in the Heisenberg group into pieces that admit an "intrinsic corona decomposition." This allows one to deduce an endpoint W1,1→L2(L1) boundedness of a certain singular integral operator from a corresponding lower-dimensional W1,2→L2(L2) boundedness. The above inequality has several applications, including that any embedding into L1 of a ball of radius n in the word metric on HZ5 incurs bi-Lipschitz distortion that is at least a constant multiple of logn. It follows that the integrality gap of the Goemans--Linial semidefinite program for the Sparsest Cut Problem on inputs of size n is at least a constant multiple of logn.
@article{arxiv.1701.00620,
title = {Vertical perimeter versus horizontal perimeter},
author = {Assaf Naor and Robert Young},
journal= {arXiv preprint arXiv:1701.00620},
year = {2018}
}