Documentation

Anderson4D.DetParametrix.Paper42_Moment.R324CollapseEntity

The general-order entity lattice pattern and its AM–GM grade #

R324CollapsePi collapses one half chain at every order. A pure-cross entity has two half chains sharing the same m cross momenta: the left chain reads them in identity order, the right chain in the order of the bijection. So the collapsed entity of the bijection σ is the zero-sum lattice sum whose summand is

W_τ(q) = ∏ᵢ ‖ρ̂(εqᵢ)‖² · P(q) · P(q∘τ), τ = σ⁻¹,

where P(q) = ∏ⱼ ⟨qⱼ+…+q_{m-1}⟩⁻² is exactly the propagator product harvested by r324Col_piChain_integral (suffix sums; on the zero-sum sector these are the paper's prefix sums up to sign). This file defines that pattern for arbitrary m and arbitrary σ, and proves what the arithmetic-geometric argument gives.

The grade the AM–GM argument gives is flat #

ab ≤ (a²+b²)/2 applied to a = P(q), b = P(q∘τ) splits every entity into two doubled patterns (r324ColLatWeight_le_doubled_avg), and the relabelling q ↦ q∘τ is a measure-preserving bijection of the lattice configuration space under which the symbol weight is invariant (r324Col_tsum_doubled_perm). Hence at every order and for every bijection the AM–GM bound is the identity entity's own bound (r324Col_tsum_latWeight_le): the argument is uniform in σ and therefore produces exactly the flat grade m-1.

That is sharp for the identity entity, and by r324CappedCross_flatGrade_over_budget it is over budget for large m. So this file also delimits precisely what is missing for the graded clause: a mechanism that is not invariant under relabelling the cross momenta, i.e. one that sees the interaction between the two suffix-sum families rather than each separately. At m = 3 the flat grade is enough (3! ≤ 2³), which is why order three closes on the calibration alone.

The lattice configuration of an entity #

def Anderson4D.r324ColExtend {m : ℕ} (q : Fin m → Z4) :
ℕ → Z4

Extend a finite momentum assignment to the sequence consumed by the chain collapse.

Equations
Instances For
    noncomputable def Anderson4D.r324ColLatSym (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (q : Fin m → Z4) :

    The symbol weight ∏ᵢ ‖ρ̂(εqᵢ)‖² of the m cross legs.

    Equations
    Instances For
      noncomputable def Anderson4D.r324ColLatProp {m : ℕ} (q : Fin m → Z4) :

      The propagator product of one half chain, read in the order q: the suffix-sum brackets delivered by r324Col_piChain_integral.

      Equations
      Instances For
        noncomputable def Anderson4D.r324ColLatWeight (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (τ : Equiv.Perm (Fin m)) (q : Fin m → Z4) :

        The lattice summand of the pure-cross entity σ at order m: symbol weights times the two half-chain propagator products, the left one in identity order and the right one in the order τ = σ⁻¹.

        Equations
        Instances For
          noncomputable def Anderson4D.r324ColLatDoubled (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (q : Fin m → Z4) :

          The doubled pattern: the same symbol weight against a squared propagator product. This is the resonant shape the arithmetic-geometric split produces.

          Equations
          Instances For
            theorem Anderson4D.r324ColLatSym_nonneg (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (q : Fin m → Z4) :
            0 ≤ r324ColLatSym ρ ε q
            theorem Anderson4D.r324ColLatWeight_nonneg (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (τ : Equiv.Perm (Fin m)) (q : Fin m → Z4) :
            0 ≤ r324ColLatWeight ρ ε τ q
            theorem Anderson4D.r324ColLatDoubled_nonneg (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (q : Fin m → Z4) :

            The arithmetic-geometric split, uniformly in the order #

            theorem Anderson4D.r324ColLatWeight_le_doubled_avg (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (τ : Equiv.Perm (Fin m)) (q : Fin m → Z4) :
            r324ColLatWeight ρ ε τ q ≤ (r324ColLatDoubled ρ ε q + r324ColLatSym ρ ε q * r324ColLatProp (q ∘ ⇑τ) ^ 2) / 2

            The AM–GM split at every order and every bijection. The mixed propagator pattern of σ is dominated by the average of the two doubled patterns: the identity one and the σ-relabelled one.

            Relabelling invariance #

            The zero-sum sector. The two external Green legs of each half chain integrate to unit mass and turn the last surviving character into the Kronecker delta q₀+…+q_{m-1} = 0 (r324Col_piChain_closed), so the collapsed entity is a sum over this sector only. It carries m-1 free momenta, which is why the flat grade is m-1 and not m.

            Equations
            Instances For
              def Anderson4D.r324ColReindex {m : ℕ} (τ : Equiv.Perm (Fin m)) :
              (Fin m → Z4) ≃ (Fin m → Z4)

              Reordering the cross momenta is a bijection of the lattice configuration space.

              Equations
              Instances For
                @[simp]
                theorem Anderson4D.r324ColReindex_apply {m : ℕ} (τ : Equiv.Perm (Fin m)) (q : Fin m → Z4) :
                (r324ColReindex τ) q = q ∘ ⇑τ

                Reordering preserves momentum conservation, so it restricts to the zero-sum sector.

                Equations
                Instances For
                  @[simp]
                  theorem Anderson4D.r324ColReindexZero_val {m : ℕ} (τ : Equiv.Perm (Fin m)) (q : R324ColZeroSum m) :
                  ↑((r324ColReindexZero τ) q) = ↑q ∘ ⇑τ
                  theorem Anderson4D.r324ColLatSym_comp (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (τ : Equiv.Perm (Fin m)) (q : Fin m → Z4) :
                  r324ColLatSym ρ ε (q ∘ ⇑τ) = r324ColLatSym ρ ε q

                  The symbol weight is symmetric in the cross momenta.

                  theorem Anderson4D.r324Col_tsum_doubled_perm (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (τ : Equiv.Perm (Fin m)) :
                  ∑' (q : R324ColZeroSum m), r324ColLatSym ρ ε ↑q * r324ColLatProp (↑q ∘ ⇑τ) ^ 2 = ∑' (q : R324ColZeroSum m), r324ColLatDoubled ρ ε ↑q

                  The relabelled doubled sum is the identity doubled sum. This is the exact reason the arithmetic-geometric argument cannot see the grading: it is invariant under permuting the cross momenta.

                  theorem Anderson4D.r324Col_summable_doubled_perm (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (τ : Equiv.Perm (Fin m)) (h : Summable fun (q : R324ColZeroSum m) => r324ColLatDoubled ρ ε ↑q) :
                  Summable fun (q : R324ColZeroSum m) => r324ColLatSym ρ ε ↑q * r324ColLatProp (↑q ∘ ⇑τ) ^ 2

                  The grade the argument delivers #

                  theorem Anderson4D.r324Col_tsum_latWeight_le (ρ : SmoothCutoff) (ε : ℝ) {m : ℕ} (τ : Equiv.Perm (Fin m)) (h : Summable fun (q : R324ColZeroSum m) => r324ColLatDoubled ρ ε ↑q) :
                  (Summable fun (q : R324ColZeroSum m) => r324ColLatWeight ρ ε τ ↑q) ∧ ∑' (q : R324ColZeroSum m), r324ColLatWeight ρ ε τ ↑q ≤ ∑' (q : R324ColZeroSum m), r324ColLatDoubled ρ ε ↑q

                  Every entity is bounded by the identity entity's doubled sum, at every order. This is the exact output of the arithmetic-geometric argument at general m and general σ: the flat grade. It is sharp for σ = id and, by r324CappedCross_flatGrade_over_budget, cannot by itself fund the graded-count clause at large m.

                  General-order analytic inputs #

                  The general-order physical→lattice collapse, as a named Prop: the physical entity integral is dominated by the zero-sum lattice sum of its own suffix-bracket pattern. This is what r324Col_piChain_integral and r324Col_piChain_closed evaluate half by half; assembling the two halves needs the covariance Fourier expansion and one Tonelli interchange between the mode sum and the 2m+4 physical integrations.

                  Equations
                  • One or more equations did not get rendered due to their size.
                  Instances For

                    The doubled window budget: the identity-pattern lattice sum, m-1 free momenta each covered by one critical quartic window, is D^m·L^{m-1}. This is the general-order form of the proved translated-window iteration r324SW_translated_window_le_log.

                    Equations
                    • One or more equations did not get rendered due to their size.
                    Instances For

                      The general-order per-entity ledger. The collapse plus the doubled window budget give the ungraded per-entity bound at every order, with the flat grade m-1 that the arithmetic-geometric argument delivers uniformly in the bijection.

                      theorem Anderson4D.r324Col_cappedCrossLedgerAt_three {ρ : SmoothCutoff} {C D : ℝ} (hC : 0 ≤ C) (hD : 0 ≤ D) (hcol : R324ColPhysicalCollapseAt ρ C 3) (hbud : R324ColDoubledBudgetAt ρ D 3) :
                      ∃ (K : ℝ), 0 ≤ K ∧ R324CappedCrossLedgerAt ρ K 3

                      Order three closes on the general machine. At m = 3 the entity count 3! = 6 is itself geometric, so the flat grade suffices and the capped cross ledger at order three follows from the two general-order inputs alone.