A conditional Green occupation lower bound from block heat
ProvedAnalyticCombinatorics.Ch8.Partitions.Erdos.green_lower_from_block_heatanalytic-combinatoricsfinite-kernel-formslean4partitions-remainder
On a finite type with decidable equality, let be a nonnegative row-stochastic kernel and a nonnegative distribution of total mass one. Assume the stated second-moment bound up to , the full block-heat lower estimate in the band, , , , and . With every bound exactly as in the contract below,
This implication does not supply the block-heat or second-moment estimates.
This support estimate carries the following complete literal source contract:
theorem green_lower_from_block_heat
(K : α → α → ℝ) (D : α → ℝ) (μ0 : α → ℝ)
(W Λ c0 V c : ℝ) (T : ℕ)
(hKnn : ∀ x z, 0 ≤ K x z) (hKsum : ∀ x, ∑ z, K x z = 1)
(hμ0nn : ∀ x, 0 ≤ μ0 x) (hμ0sum : ∑ x, μ0 x = 1)
(hM2 : ∀ m, m ≤ T / 2 →
∑ x, distPow K μ0 m x * (D x) ^ 2 ≤ V * ((m : ℝ) + 1))
(hΛpos : 0 < Λ) (hΛ : 2 * V ≤ Λ ^ 2)
(hc0 : 0 ≤ c0) (hT : 16 ≤ T) (hc : c ≤ c0 / 8)
(hblock : ∀ m, m ≤ T / 2 → 1 ≤ m → ∀ x,
|D x| ≤ Λ * Real.sqrt ((m : ℝ) + 1) →
c0 / Real.sqrt ((m : ℝ) + 1)
≤ ∑ y, distPow K (fun z => if z = x then 1 else 0) m y *
(if |D y| ≤ W then (1 : ℝ) else 0)) :
c * Real.sqrt (T : ℝ)
≤ ∑ t ∈ Finset.range T,
∑ x, distPow K μ0 t x *
(if |D x| ≤ W then (1 : ℝ) else 0)
Preamble
import Definitions.Def_AnalyticCombinatorics_Ch8_Partitions_TanakaTelescope
import Mathlib
set_option autoImplicit true
set_option relaxedAutoImplicit false
set_option maxSynthPendingDepth 3
/-!
# Green lower bridge from block heat + second-moment tightness (renewal route B, Layer 1)
This is the *deterministic finite-sum* bridge of the two-layer split (ChatGPT R8):
given a per-block heat lower bound `hblock` (`c0/√(m+1) ≤` window mass from a tight start) and a
second-moment tightness bound `hM2` (`∑ distPow·D² ≤ V(m+1)`), the cumulative window occupation over
a horizon `T` is `≥ (c0/8)·√T`. No probability/measure theory — pure `Finset` algebra:
* `good_mass_ge_half` — Chebyshev/Markov: ≥ ½ of the law sits in `{|D| ≤ Λ√(m+1)}`.
* `distPow_expect_comp` — Chapman–Kolmogorov in expectation form (induction on the recursive `distPow`).
* `even_time_windowMass_lower` — one even-time block: feed `hblock` from each good start.
* `sum_Icc_one_div_sqrt_succ_ge` + `sqrt_T_quarter_le_sqrt_half_minus_one` — the `∑ 1/√(m+1) ≥ √T/4`
telescope (no integrals).
* `green_lower_from_block_heat` — the assembled bridge, constant `c ≤ c0/8`, needs `16 ≤ T`.
The remaining wall is the *Layer 2* input `hblock` itself (`erdos_rankdiff_block_heat_lower`); this file
discharges everything around it. (Bridge authored via ChatGPT ac2 R8, verified against the repo
`distPow` API and rebuilt clean-3.)
-/
open Finset
open AnalyticCombinatorics AnalyticCombinatorics.Ch8 AnalyticCombinatorics.Ch8.Partitions AnalyticCombinatorics.Ch8.Partitions.Erdos
variable {α : Type*} [Fintype α] [DecidableEq α]
Formal statement
theorem AnalyticCombinatorics.Ch8.Partitions.Erdos.green_lower_from_block_heat
(K : α → α → ℝ) (D : α → ℝ) (μ0 : α → ℝ)
(W Λ c0 V c : ℝ) (T : ℕ)
(hKnn : ∀ x z, 0 ≤ K x z) (hKsum : ∀ x, ∑ z, K x z = 1)
(hμ0nn : ∀ x, 0 ≤ μ0 x) (hμ0sum : ∑ x, μ0 x = 1)
(hM2 : ∀ m, m ≤ T / 2 →
∑ x, distPow K μ0 m x * (D x) ^ 2 ≤ V * ((m : ℝ) + 1))
(hΛpos : 0 < Λ) (hΛ : 2 * V ≤ Λ ^ 2)
(hc0 : 0 ≤ c0) (hT : 16 ≤ T) (hc : c ≤ c0 / 8)
(hblock : ∀ m, m ≤ T / 2 → 1 ≤ m → ∀ x,
|D x| ≤ Λ * Real.sqrt ((m : ℝ) + 1) →
c0 / Real.sqrt ((m : ℝ) + 1)
≤ ∑ y, distPow K (fun z => if z = x then 1 else 0) m y *
(if |D y| ≤ W then (1 : ℝ) else 0)) :
c * Real.sqrt (T : ℝ)
≤ ∑ t ∈ Finset.range T,
∑ x, distPow K μ0 t x *
(if |D x| ≤ W then (1 : ℝ) else 0) := by sorrySource
Human review
Confirmed by the mission captain (proposal self-audit).