Documentation

MathlibNt.SieveTheory.LinearSieve.Suzuki.SuzukiFiniteContinuousLayers

Suzuki's finite continuous layers: source-correct normalized recursion #

This file records Suzuki, §9, (9.1)--(9.2). The decisive point is that the integrand in (9.2) is the preceding normalized layer f_{n-1}(t - 1), not its unnormalized numerator.

The normalized layer is therefore defined first by structural recursion. The numerator is then defined as the right-hand side of (9.1) or (9.2), so that the source equations are available without division. Their normalization bridge is valid for every s; cancellation in the opposite direction requires a nonzero normalizing factor (below we expose the natural hypothesis 0 < s).

Normalized form of (9.1).

theorem MathlibNt.SieveTheory.SuzukiFiniteContinuousLayers.suzukiLayer_succ_succ (κ β s : ) (n : ) :
suzukiLayer κ β (n + 2) s = (s ^ κ)⁻¹ * (t : ) in recursionLower β s (n + 2)..β + (n + 2), suzukiLayer κ β (n + 1) (t - 1) * dPowDensity κ t

Normalized form of (9.2), with the preceding normalized layer in the integrand.

Unnormalized source right-hand side in (9.1).

theorem MathlibNt.SieveTheory.SuzukiFiniteContinuousLayers.suzukiLayerNumerator_succ_succ (κ β s : ) (n : ) :
suzukiLayerNumerator κ β (n + 2) s = (t : ) in recursionLower β s (n + 2)..β + (n + 2), suzukiLayer κ β (n + 1) (t - 1) * dPowDensity κ t

Source-correct unnormalized right-hand side in (9.2).

The division-form bridge is total, including at s = 0.

Cancellation is valid whenever the normalizing factor is nonzero.

The source equation s^κ fₙ(s) = RHSₙ(s) after legal cancellation on the positive Suzuki domain.

@[simp]

At the excluded point s = 0, positive κ makes the normalized definition zero. No theorem identifies its numerator with 0: multiplying by 0^κ cannot recover an arbitrary source right-hand side.

@[simp]

Proposition 9.2(i), at the level of the source RHS.

Proposition 9.2(i), for a normalized finite layer.

theorem MathlibNt.SieveTheory.SuzukiFiniteContinuousLayers.suzukiLayerNumerator_succ_succ_nonneg {κ β s : } {n : } ( : 0 κ) ( : 1 β) (hprev : tSet.Icc (recursionLower β s (n + 2)) (β + (n + 2)), 0 suzukiLayer κ β (n + 1) (t - 1)) :
0 suzukiLayerNumerator κ β (n + 2) s

One source-correct recursive positivity step: the induction hypothesis is on the preceding normalized layer.