Spatial correlation matrices tend to the two stated endpoints
ProvedQAlgorithms.Papers.Independent.Advance50.P39.correlationMatrixLimitsLet and let be the intermolecular distances for distinct sites, with and when . For positive correlation radius , define . Entrywise,
These are the uncorrelated and perfectly correlated matrix limits. The first is a one-sided limit; assigning the totalized real quotient a value at would not represent the physical endpoint.
import Definitions.Def_QAlgorithms_Papers_Independent_Advance50_P39
noncomputable section
open MeasureTheory Filter
open scoped BigOperators Matrix ComplexOrder Topology Matrix.Norms.L2Operator
open QAlgorithms.Papers.Independent.Dependency50.P16 (sandwich semigroupEvolution)
namespace QAlgorithms.Papers.Independent.Advance50.P39
attribute [local instance] Classical.propDecidable Classical.decEq
/-- PDF pp.4--5, §IV: uncorrelated and perfectly correlated limits.
At radius zero use the right-hand limit, never real division by zero. -/
theorem correlationMatrixLimits {n : ℕ} (hn : 0 < n)
(distance : Fin n → Fin n → ℝ) (hdiag : ∀ i, distance i i = 0)
(hoffdiag : ∀ i j, i ≠ j → 0 < distance i j) :
(∀ i j, Tendsto (fun radius => spatialCorrelation (distance i j) radius)
(nhdsWithin 0 (Set.Ioi 0)) (nhds (if i = j then 1 else 0))) ∧
(∀ i j, Tendsto (fun radius => spatialCorrelation (distance i j) radius)
atTop (nhds 1)) := by sorry
end QAlgorithms.Papers.Independent.Advance50.P39
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For every natural number and every real-valued array , indexed by , suppose that for every and that whenever . Then, for every pair ,
The first limit approaches zero through strictly positive radii. No symmetry or triangle inequality is assumed for ; when , the off-diagonal positivity hypothesis is vacuous.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.