Documentation

Anderson4D.DetParametrix.Paper42_Moment.R324ScaledCentralBudget

The correctly scaled central bracket, and where it comes from #

R324AnchorCentralBudget asks the anchor lattice series to carry the ε-free bracket ⟨‖α+β‖⟩⁻⁸. That is stronger than clause B needs: clause B's own weight is

r324CMBracketWeight ε α β = ε⁻⁸⟨α⟩⁻⁴⟨β⟩⁻⁴ · ⟨ε‖freq(α+β)‖⟩⁻⁸,

whose bracket is ε-scaled and which carries a spare ε⁻⁸. The ε-free demand exceeds the honest target by at least ε⁻⁸, and that excess is exactly the shortfall which made the half-symbol route look impossible (r324Central_epsScale_gap).

At the correct scaling the half-symbol route works, unconditionally:

The output is R324ScaledAnchorCentralBudget, the honest replacement of R324AnchorCentralBudget, with the ε⁻⁸ slack still unspent.

One power of the symbol carries the eighth-order bracket #

theorem Anderson4D.r324Scaled_exists_symbolBracket (ρ : SmoothCutoff) :
∃ (C : ℝ), 0 < C ∧ ∀ {ε : ℝ}, 0 < ε → ∀ (k : Z4), ‖ρ.symbol ε k‖ ≤ C * eighthOrderFrequencyDecay (ε * ‖z4EuclideanFrequency k‖)

The single-power symbol bracket. The order-8 Schwartz decay of ρ̂ gives ‖ρ̂(εk)‖ ≤ C⟨ε‖k‖⟩⁻⁸ — a whole eighth-order bracket per key from one of the two symbol factors, the 2π Fourier normalization absorbed into C. This is the half-symbol split of r324HdetAssembly_exists_halfSymbol_sq_bracket in the form the product bracket consumes.

The half-symbol weight #

noncomputable def Anderson4D.r324ScaledHalfSym (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (q : Fin m → Z4) :

The remaining half of the symbol weight: one power of ρ̂(εqᵢ) per key instead of two. r324ColLatSym is its square.

Equations
Instances For
    theorem Anderson4D.r324ScaledHalfSym_nonneg (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (q : Fin m → Z4) :

    The product bracket on the conservation sector #

    theorem Anderson4D.r324Scaled_halfSym_le_bracket (ρ : SmoothCutoff) {C : ℝ} (hC : 0 ≤ C) (hsym : ∀ {ε : ℝ}, 0 < ε → ∀ (k : Z4), ‖ρ.symbol ε k‖ ≤ C * eighthOrderFrequencyDecay (ε * ‖z4EuclideanFrequency k‖)) {ε : ℝ} (hε : 0 < ε) {m : ℕ} (hm : 1 ≤ m) (γ : Z4) (q : Fin m → Z4) (hq : ∑ i : Fin m, q i = γ) :

    The central bracket, harvested. On the momentum sector ∑ᵢ qᵢ = γ the half-symbol weight is bounded by Cᵐ·m⁴ times the eighth-order bracket at the conserved mode, at the ε-scaled frequency. No zone split, no endpoint trade: the product bracket r324HdetAssembly_prod_eighthDecay_le and momentum conservation do all the work.

    The half-symbol anchor weight and the window budget #

    noncomputable def Anderson4D.r324ScaledHalfAnchorWeight (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (α β : Z4) (aL aR : ℕ) (τ : Equiv.Perm (Fin m)) (q : Fin m → Z4) :

    The anchor lattice summand with the symbol at half power: what is left after the product bracket has been harvested.

    Equations
    Instances For
      theorem Anderson4D.r324ScaledHalfAnchorWeight_nonneg (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (α β : Z4) (aL aR : ℕ) (τ : Equiv.Perm (Fin m)) (q : Fin m → Z4) :
      0 ≤ r324ScaledHalfAnchorWeight ρ ε α β aL aR τ q

      The window budget, at half symbol power. Exactly the proved window-budget shape — R324ColGradedBudgetAt ρ D m at the flat grade m-1, equivalently R324AnchorCentralBudget — with two changes: the lattice symbol weight sits at half power (the other half has been spent on the central bracket) and no frequency bracket is demanded on the right.

      Equations
      • One or more equations did not get rendered due to their size.
      Instances For
        theorem Anderson4D.r324Scaled_anchorLatWeight_le_half (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (α β : Z4) (aL aR : ℕ) (τ : Equiv.Perm (Fin m)) (q : Fin m → Z4) :
        r324AnchorLatWeight ρ ε α β aL aR τ q ≤ r324ScaledHalfAnchorWeight ρ ε α β aL aR τ q

        How the half-symbol budget relates to the proved one. Since ‖ρ̂(εk)‖ ≤ 1, the half-symbol weight dominates the proved anchor lattice weight pointwise. So R324ScaledHalfWindowBudget is exactly the proved window budget with the symbol at half power — the same statement about the same Schwartz-decaying symbol, and the only price paid for the central bracket.

        The correctly scaled central lattice budget. The honest replacement of R324AnchorCentralBudget: the bracket at the conserved mode is the ε-scaled one, which is all r324CMBracketWeight demands, and the ε⁻⁸ endpoint slack is not spent here at all.

        Equations
        • One or more equations did not get rendered due to their size.
        Instances For
          theorem Anderson4D.r324Scaled_anchorLatWeight_le (ρ : SmoothCutoff) {C : ℝ} (hC : 0 ≤ C) (hsym : ∀ {ε : ℝ}, 0 < ε → ∀ (k : Z4), ‖ρ.symbol ε k‖ ≤ C * eighthOrderFrequencyDecay (ε * ‖z4EuclideanFrequency k‖)) {ε : ℝ} (hε : 0 < ε) {m : ℕ} (hm : 1 ≤ m) (α β : Z4) (bL bR : ℕ) (τ : Equiv.Perm (Fin m)) (q : Fin m → Z4) (hq : ∑ i : Fin m, q i = -(α + β)) :
          r324AnchorLatWeight ρ ε α β bL bR τ q ≤ C ^ m * ↑m ^ 4 * eighthOrderFrequencyDecay (ε * ‖z4EuclideanFrequency (α + β)‖) * r324ScaledHalfAnchorWeight ρ ε α β bL bR τ q

          The pointwise harvest. On the conservation sector every anchor lattice summand is the scaled central bracket times its half-symbol companion.

          The scaled central budget, proved #

          The correctly scaled central budget, from the window budget alone. This is the step that r324Central_epsScale_gap shows is impossible for the ε-free target and which is unconditional here: the covariance symbols only ever see εk, and the ε-scaled bracket is exactly what they produce. The ε⁻⁸ slack of r324CMBracketWeight is never touched.