Lean source theorem: canonical_tail_ratio_quantitative
ProvedErdosProblems.Erdos243.PaperCompleteR7.canonical_tail_ratio_quantitativeerdos-243support-lemma
For a strictly increasing positive integer sequence with rational reciprocal sum and a(n+1)/a(n)² tending to one, the canonical numerator's ratio differs from a(n)²/a(n+1) by at most 16/a(n) eventually, and a(n) has a doubly exponential lower bound exp(c·2ⁿ) for some c>0.
Preamble
import Definitions.Def_ErdosProblems_Erdos243_PaperCompleteR7_RealTail import Definitions.Def_ErdosProblems_Erdos243_PaperCompleteR7_CanonicalState import Mathlib.Algebra.Order.BigOperators.GroupWithZero.Finset import Mathlib.Analysis.SpecialFunctions.Log.Basic import Mathlib.Analysis.SpecificLimits.Basic import Mathlib.Analysis.SpecificLimits.Normed import Mathlib.Data.Fin.Pigeonhole import Mathlib.Data.Nat.ChineseRemainder import Mathlib.Data.Nat.Find import Mathlib.Data.Nat.Prime.Basic import Mathlib.Data.ZMod.Basic import Mathlib.Order.Filter.AtTopBot.Basic import Mathlib.Tactic import Mathlib.Tactic.Ring import Mathlib.Topology.Algebra.InfiniteSum.NatInt /-! # The quantitative canonical-tail lemma Compiled candidates for the whole long-record `res:tailratio`. The O(1/a_n) conclusion has an explicit eventual constant 16. The growth bound is first proved as an exact binary-power inequality, and then converted to the exponential notation of the paper. -/ open Filter open ErdosProblems.Erdos243.PaperCompleteR7
Formal statement
theorem ErdosProblems.Erdos243.PaperCompleteR7.canonical_tail_ratio_quantitative
(a : ℕ → ℕ) (ha : StrictMono a) (hpos : ∀ n, 0 < a n)
(p : ℤ) (q : ℕ) (hq : 0 < q)
(hs : HasSum (fun n ↦ 1 / (a n : ℝ)) ((p : ℝ) / (q : ℝ)))
(hgrowth : Tendsto (fun n ↦ (a (n + 1) : ℝ) / (a n : ℝ) ^ 2)
atTop (nhds 1)) :
let C := canonicalNaturalNumerator a p q
(∃ N, ∀ n, N ≤ n →
|(C (n + 1) : ℝ) / (C n : ℝ) -
(a n : ℝ) ^ 2 / (a (n + 1) : ℝ)| ≤ 16 / (a n : ℝ)) ∧
(∃ c : ℝ, 0 < c ∧ ∃ N, ∀ n, N ≤ n →
Real.exp (c * (2 : ℝ) ^ n) ≤ (a n : ℝ)) := by sorry
Source
Lean source (Apache-2.0): https://github.com/wcook04/plectis-erdos-lean/blob/4fe59e0191169606942baae0949365070b7f419c/ErdosProblems/Erdos243/PaperCompleteR7/QuantitativeTail.lean#L201-L234
Related paper by Will Cook (CC-BY-4.0): https://github.com/wcook04/plectis-erdos/blob/6917e15ec4abc2623512254da93221e446eeb707/paper/243/erdos-243-reciprocal-tail-rigidity.tex#L1-L58
Paper's AI-assistance disclosure: https://github.com/wcook04/plectis-erdos/blob/6917e15ec4abc2623512254da93221e446eeb707/paper/paper-house-style.sty#L180-L188
Original Erdős problem and Koizumi prior work are distinguished in the paper: https://github.com/wcook04/plectis-erdos/blob/6917e15ec4abc2623512254da93221e446eeb707/paper/243/erdos-243-reciprocal-tail-rigidity.tex#L57-L110