Documentation

LeanPool.Superdiffusion.Algsuperdiff.Section4.Provider.ExcessDecay.EdBridgeStepFourDatumSplit

EdBridgeStepFourDatumSplit #

Ported from scottnarmstrong/Superdiffusion at 7f20476385d4af418f110303085ad0740c9eba46. Module imports, support namespace and local option directives adapted for Lean Pool.

The Step-4 bridge with the printed datum leg threaded through #

EdBridgeStepFourGeneral.excessDecay_stepFour_slot_general runs the residue chain against the join and delivers the Step-4 slot at every centre. Its join input is the two-leg one-step; this module runs the same chain against the anchored one-step with the printed datum leg, whose boundary disjunct carries its binders honestly and whose conclusion carries the printed third leg

   C_t(d) · (3^{n-k})^{1/2} · (boundaryDatumLegConst d C k · K_h) .

The additive ride #

One risk has to be excluded: the datum leg's fixed-k factor 3^{3k/2} would be fatal if the leg ever entered the bridge's absorption inequality — the k₀ selection of exists_edBridgeStepGen, which is what forces k large enough that the contraction is ≤ ½·3^{-(k+1)/4}. It does not. The machine witness is the shape of the generalized recombination edBridge_recombine_datum: the leg appears as a single free real Leg which

Since k₀ is chosen by exists_edBridgeStepGen from Cs alone — a statement this module does not touch and whose proof mentions no datum object — the 3^{3k/2} cannot interact with the selection. Concretely: k₀ is fixed first, k ≥ k₀ is arbitrary, and the leg's constant is then evaluated at that k. The leg is a δ-side summand, not an E-side coefficient.

The consumption shape #

edBridgeDatumLeg_eq rewrites the leg into's A^h shape exactly: at the slot's own index j = n+1,

   C_t(d)·(3^{n-k})^{1/2}·(C_bd(d,C,k)·K_h)  =  edBridgeDatumLegConst d C k · (3^{j/2} · K_h) ,

a constant times 3^{j/2} times the same datum K_h the budget's A^h_j leg carries. So the leg is absorbed by widening A^h's constant, not by touching the ε-free flat slot F (which is the anchor's K_hinf leg). measured this; the identity proves it.

References #

1. The four-leg recombination: the datum leg rides additively #

theorem Algsuperdiff.Section4.Provider.ExcessDecay.edBridge_recombine_datum {Elhs Acon Ejm1 Ej pj Big Rhat P Crem Sq B₁ B₂ th₁ th₂ th eps Wd Leg : ℝ} (hCrem : 0 ≤ Crem) (hSq : 0 ≤ Sq) (hP : 0 ≤ P) (hEj : 0 ≤ Ej) (hmain : Elhs ≤ Acon * Ejm1 + Crem * (P * (Sq * Big)) + Leg) (hBig : Big ≤ B₁ * Ej + B₂ * pj + Rhat) (hcon : Acon * Ejm1 ≤ th₁ * Ej) (hgate : Crem * (P * (Sq * B₁)) ≤ th₂) (heps : Crem * (P * (Sq * B₂)) = eps) (hWd : Crem * (P * (Sq * Rhat)) = Wd) (hth : th₁ + th₂ ≤ th) :
Elhs ≤ th * Ej + eps * pj + (Wd + Leg)

EdBridgeStepFour.edBridge_recombine with a fourth, purely additive leg.

Leg occurs only in hmain and in the conclusion, in both cases as a bare additive summand. It touches neither the δ-gate hgate nor the contraction absorption hcon nor the step budget hth — this is the machine form's "the datum leg rides additively through the k₀ selection".

2. The datum leg and its A^h shape #

noncomputable def Algsuperdiff.Section4.Provider.ExcessDecay.edBridgeDatumLeg (d : ℕ) (Cs : ℝ) (k : ℕ) (n : ℤ) (Kh : ℝ) :

The datum leg the bridge's slot carries, at the one-step's own scale n: C_t(d) · (3^{n-k})^{1/2} · (C_bd(d,Cs,k) · K_h).

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

    The datum leg's constant once the slot's own 3^{j/2} is factored out (j = n+1): C_t(d) · (3^{-(k+1)})^{1/2} · C_bd(d,Cs,k).

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

      (3^m)^{1/2} = 3^{m/2} for an integer power.

      theorem Algsuperdiff.Section4.Provider.ExcessDecay.edBridgeDatumLeg_eq (d : ℕ) (Cs : ℝ) (k : ℕ) (n : ℤ) (Kh : ℝ) :
      edBridgeDatumLeg d Cs k n Kh = edBridgeDatumLegConst d Cs k * (3 ^ (↑(n + 1) / 2) * Kh)

      The datum leg is exactly's A^h_j shape at j = n+1.

      edBridgeDatumLeg d Cs k n K_h = edBridgeDatumLegConst d Cs k · (3^{j/2} · K_h) — a constant times 3^{j/2} times the same datum K_h the budget's A^h leg carries. So the new leg is absorbed by widening A^h's constant, not by touching the ε-free flat slot.