Definition 1.6 — the Integrated event
DefinitionSpatialSorting_Supplement_IntegratedLet be a density vector and positive reals. Write for the number of free sites. A configuration (colours on , the collar wired to colour , fixed colour counts) belongs to when the following holds:
for every region (not necessarily connected) of boundary length , and every colour ,
The Integrated event says that colours are spread out: every low-perimeter region carries, up to a macroscopic error , its fair share of each colour. It is the high-temperature counterpart of the Sorted event.
Formalization Note The definition is the predicate on configurations; membership in is imposed where the event is measured (integratedProb), as for the base Sorted. The region test is a real comparison, both bounds are inclusive, the empty region and are included. The predicate is defined for all real ; positivity is a hypothesis of the theorem.
import Mathlib
import Definitions.Def_SpatialSorting_I
import Definitions.Def_SpatialSorting_N
import Definitions.Def_SpatialSorting_Coloring
import Definitions.Def_SpatialSorting_extendBy
import Definitions.Def_SpatialSorting_nV
import Definitions.Def_SpatialSorting_Supplement_bdryLen
open Classical Filter Topology
namespace SpatialSorting.Supplement
/-- Definition 1.6 (the Integrated event): a configuration `σ` is in `Integrated(c, δ)` when, for every
region `R ⊆ I_L` of boundary length at most `c √|R|` and every colour `i ∈ {0, …, q − 1}`, the number of
sites of `R` with colour `i` lies between `ρ_i |R| − δ N_L` and `ρ_i |R| + δ N_L`.
Calvert, Oh, Randall, *Spatial Sorting and the Geometry of Lattice Spin Systems at Fixed Magnetization*, STOC 2027 submission manuscript (sha256 4d4d4b7d…7fae), Definition 1.6, p. 9 (PDF p. 10).
Formalization Notes:
- `Integrated(c, δ)` is a subset of `Ω_{Λ^L,ρ}`; here it is the predicate, and membership in
`omegaRho ρ L` is imposed where the event is measured (`integratedProb`), as for the base `Sorted`.
- Regions are arbitrary (not necessarily connected) subsets of `I_L`, including `∅` and `I_L`; the
test `|∂_E R| ≤ c √|R|` is a real comparison, and both bounds of the count are inclusive.
- The colour count is the base `nV R (extendBy 0 (I L) σ) i`; `N_L` is the base `N L = |I_L|`.
- The source takes `c, δ > 0`; the predicate is defined for all real `c, δ` and the theorem assumes
positivity. -/
def Integrated {q : ℕ} [NeZero q] (ρ : Fin q → ℝ) (L : ℕ) (c δ : ℝ) (σ : Coloring q L) : Prop :=
∀ R : Finset (ℤ × ℤ), R ⊆ I L → (bdryLen L R : ℝ) ≤ c * Real.sqrt (R.card : ℝ) →
∀ i : Fin q,
ρ i * (R.card : ℝ) - δ * (N L : ℝ) ≤ (nV R (extendBy 0 (I L) σ) i : ℝ) ∧
(nV R (extendBy 0 (I L) σ) i : ℝ) ≤ ρ i * (R.card : ℝ) + δ * (N L : ℝ)
end SpatialSorting.Supplement