Bunkbed connectivity on Hollom's hypergraph depends only on the two level profiles
ProvedBunkbedTM.profile_indicatorIn the Wierman--Ziff five-state model on Hollom's -uniform hypergraph, the difference between the cross-level and the same-level connection probability is a sum over the pairs of level configurations. This statement performs the two reductions that make that sum finite to work with.
First, the two levels are conditionally independent: each hyperedge draws its state independently on each level, and the posts over the transversal set are always present. So the sum factors into a double sum over a level- configuration and a level- configuration , weighted by .
Second — and this is the combinatorial heart — whether vertex on level is connected to vertex on level depends on and only through the partitions they induce on the five special vertices : the two endpoints of the connection event together with the transversal set. Nothing else about the configuration matters, because a walk can change level only at a post, and the posts stand over , which is part of the recorded boundary. Writing prof for that induced partition and glueConn for the resulting glued connectivity, the indicator of the connection event is exactly .
import Definitions.Def_BunkbedTransferTables open Bunkbed Bunkbed.Hyper Bunkbed.TM SimpleGraph Finset
theorem BunkbedTM.profile_indicator (P : WZ → ℚ) :
wzProb hollomTriple hollomT P ((0 : Fin 10), (0 : Fin 2)) ((9 : Fin 10), (1 : Fin 2))
- wzProb hollomTriple hollomT P ((0 : Fin 10), (0 : Fin 2)) ((9 : Fin 10), (0 : Fin 2))
= ∑ c₀ : Fin 6 → WZ, ∑ c₁ : Fin 6 → WZ,
((if glueConn (prof c₀) (prof c₁) 1 then (1 : ℚ) else 0)
- (if glueConn (prof c₀) (prof c₁) 0 then (1 : ℚ) else 0))
* ((∏ i, P (c₀ i)) * (∏ i, P (c₁ i))) := by sorry