OAI/NumberTheory/Ostmann/Characters/AnchorNonbulkBudget.lean

GitHub source file

4 declarations, 0 contain sorry, 0 declared as axioms

Nonbulk normalization at the manuscript's exponential bulk scale

  1. theorem OAI.Ostmann.eventual_anchor_nonbulk_cost line 12
  2. theorem OAI.Ostmann.eventual_scheduled_anchor_nonbulk_cost line 39
  3. theorem OAI.Ostmann.eventual_active_anchor_nonbulk_cost line 76

    Source documentation

    The actual H block only pays for original active nonword slots. Negative word copies and the filler never enter this count.

    theorem eventual_active_anchor_nonbulk_cost (n B : ℕ) (c ε z : ℝ)
        (hc : 0 ≤ c) (hε : 0 < ε) (hz : 0 < z)
        (hbudget : 4 * c * B ≤ ε * z) :
        ∀ᶠ L : ℝ in atTop, ∀ {I : Type*} [Fintype I]
          (role : I → CopyScheduleRole) (m : ℕ),
          z * L / 2 ≤ (m : ℝ) →
          Fintype.card {i : I // role i ≠ .word ∧ role i ≠ .outside} ≤ B →
          ∀ mass : CopyScheduleH role n → ℝ,
          (∀ h : ScheduledNonbulkH role n, Real.exp (-c * L) ≤ mass h.val) →
          (Fintype.card (ScheduledNonbulkH role n)).factorial *
            (∏ h : ScheduledNonbulkH role n, (mass h.val)⁻¹) ≤
              Real.exp (ε * (2 ^ n : ℕ) * (m : ℝ)) := by
    …

    Mentions: OAI.Ostmann.CopyScheduleRole, OAI.Ostmann.CopyScheduleH, OAI.Ostmann.ScheduledNonbulkH

  4. theorem OAI.Ostmann.eventual_anchor_role_budget line 125

    Source documentation

    The required normalization budget is available at every sufficiently large depth, from the actual cell-role bound of Section 4.

    theorem eventual_anchor_role_budget (R c ε : ℝ) (hc : 0 < c) (hε : 0 < ε) :
        ∀ᶠ k : ℕ in atTop, ∀ B : ℕ, (B : ℝ) ≤ cellRoleCountBound R k →
          4 * c * B ≤ ε * cellRoleScale k := by
    …

    Mentions: OAI.Ostmann.cellRoleCountBound, OAI.Ostmann.cellRoleScale