OAI/GroupTheory/PolycyclicRecognition/Analysis/NegativePathWords.lean
3 declarations, 0 contain sorry, 0 declared as axioms
Unit-step words remain in the negative halfspace
This is the concrete subdivision needed for quotient-boundary independence. The logarithmic mesh error is absorbed into half of the negative height gap.
- theorem
OAI.PolycyclicRecognition.one_parameter_nat_dist_leline 25Source documentation
Integer points of an actual one-parameter subgroup lie in a linear ball. This supplies the absolute radius of the negative-height ray pivots.
theorem one_parameter_nat_dist_le (d : ℝ → G) (hzero : d 0 = 1) (hmul : ∀ s t, d (s + t) = d s * d t) (n : ℕ) : dist 1 (d n) ≤ (n : ℝ) * dist 1 (d 1) := by … - theorem
OAI.PolycyclicRecognition.exists_negative_word_prefix_chainsline 38Source documentation
Retain the actual prefixes of an eventually specified unit-step word. Their absolute radius follows by adding the radius of the starting point.
theorem exists_negative_word_prefix_chains (height : G → ℝ) (base target : ℕ → G) {C D gap : ℝ} {q : ℕ} (hbase : ∀ n : ℕ, dist 1 (base n) ≤ D * n) (hwords : ∀ᶠ n : ℕ in atTop, ∃ w : List G, (∀ g ∈ w, g ∈ closedBall (1 : G) 1) ∧ w.prod = (base n)⁻¹ * target n ∧ (w.length : ℝ) ≤ C * (1 + n) ^ q ∧ ∀ i : ℕ, height (base n * (w.take i).prod) ≤ -gap * n ∧ dist (base n) (base n * (w.take i).prod) ≤ C * n) : ∃ (m : ℕ → ℕ) (z : ℕ → ℕ → G), (∀ n, z n 0 = base n) ∧ (∀ᶠ n in atTop, z n (m n) = target n) ∧ (∀ᶠ n in atTop, (m n : ℝ) ≤ C * (1 + n) ^ q) ∧ … - theorem
OAI.PolycyclicRecognition.eventually_negative_path_unit_wordsline 92Source documentation
Subdivision can start at any point of the actual negative-height path. Every resulting prefix stays nearby, has negative linear height, and the number of unit steps is bounded by a fixed quadratic.
theorem eventually_negative_path_unit_words (hpath : ∀ g : G, ∃ γ : ℝ → G, γ 0 = 1 ∧ γ 1 = g ∧ ∀ t ∈ Icc (0 : ℝ) 1, ∀ u ∈ Icc (0 : ℝ) 1, dist (γ t) (γ u) ≤ dist 1 g * |t - u|) (height : G → ℝ) {K C a : ℝ} (hK : 0 ≤ K) (hC : 0 ≤ C) (ha : 0 < a) (hheight : ∀ x y, |height x - height y| ≤ K * dist x y) (source target : ℕ → G) (hγ : ∀ᶠ n : ℕ in atTop, ∃ γ : Path (source n) (target n), PathMetricBound γ (C * n) (C * (1 + Real.log (2 + n))) ∧ ∀ t, height (γ t) ≤ -a * n) : ∀ᶠ n : ℕ in atTop, ∃ w : List G, (∀ x ∈ w, x ∈ closedBall (1 : G) 1) ∧ …