Documentation

MathlibNt.SieveTheory.Switching.SuzukiPrimeSums

Dimension-one prime sums and finite Abel identities #

Local product bounds, finite prime nodes, and Abel summation establish the dimension-one prime-sum comparison of Suzuki's Lemma 8.6.

All declarations retain the MathlibNt.SieveTheory.SwitchingPrinciple namespace.

The standard dimension-one local-product condition V(z₁)/V(z₂) ≤ (log z₂/log z₁)(1 + K/log z₁). The product is restricted to the actual finite sieving primes, so omitted Goldbach primes dividing N are handled without changing the abstract hypothesis.

Equations
Instances For

    The finite-support version of Suzuki's ratio V(w) / V(z), restricted to the actual sieving primes.

    Equations
    Instances For

      For nonnegative real cutoffs, the supported primes in [w,z) are exactly those in the natural half-open interval [⌈w⌉₊,⌈z⌉₊).

      Filter-form corollary of the exact real/natural carrier bridge.

      Suzuki equation (8.1) in dimension one follows directly from the generic local-product condition; no prime-specific PNT or Mertens theorem is used.

      The exact finite prime sum on the left side of Suzuki Lemma 8.6 in sieve dimension one. The suffix product is the finite ratio V(p)/V(z).

      Equations
      Instances For

        Supported primes below the natural cutoff z.

        Equations
        Instances For

          The finite suffix Euler ratio over supported primes n ≤ q < z.

          Equations
          Instances For

            Exact bridge from real cutoffs to natural ceiling cutoffs. In particular, an integral right endpoint retains the strict condition p < z.

            At a natural right endpoint the ceiling bridge has no off-by-one shift.

            Exact difference of consecutive finite suffix Euler products. The factor is R n, not R (n+1).

            The Suzuki prime sum is exactly the finite Abel difference sum; unsupported indices contribute zero by the suffix-product difference identity.

            theorem MathlibNt.SieveTheory.SwitchingPrinciple.finiteAbelIdentity (R F : ) (w z : ) (hwz : w z) :
            nFinset.Ico w z, (R n - R (n + 1)) * F n = R w * F w - R z * F z + nFinset.Ico (w + 1) (z + 1), R n * (F n - F (n - 1))

            Finite Abel (summation-by-parts) identity on a natural interval. This boundary convention remains valid for the empty interval w = z.

            theorem MathlibNt.SieveTheory.SwitchingPrinciple.suzukiPrimeSumNat_abelIdentity (S : BoundingSieve) (w z : ) (hwz : w z) (F : ) :
            suzukiPrimeSumNat S w z F = suzukiSuffixRatio S z w * F w - suzukiSuffixRatio S z z * F z + nFinset.Ico (w + 1) (z + 1), suzukiSuffixRatio S z n * (F n - F (n - 1))

            Suzuki's supported-prime sum in Abel boundary-plus-increment form.

            At integral cutoffs, the current Suzuki prime-sum interface is exactly the natural supported-prime sum used by the finite Abel identity.

            theorem MathlibNt.SieveTheory.SwitchingPrinciple.suzukiLemmaEightSixPrimeSum_abelIdentity (S : BoundingSieve) (D : ) (w z : ) (hwz : w z) (H : ) :
            suzukiLemmaEightSixPrimeSum S D (↑w) (↑z) H = suzukiSuffixRatio S z w * H (Real.log D / Real.log w) - suzukiSuffixRatio S z z * H (Real.log D / Real.log z) + nFinset.Ico (w + 1) (z + 1), suzukiSuffixRatio S z n * (H (Real.log D / Real.log n) - H (Real.log D / Real.log ↑(n - 1)))

            Thus the exact Suzuki prime sum at integral cutoffs has the Abel boundary-plus-increment expansion.

            The generic local-product input specializes exactly to the natural suffix ratio used by suzukiFiniteAbel.

            Equation (8.1) at natural endpoints, obtained without PNT or Mertens.

            theorem MathlibNt.SieveTheory.SwitchingPrinciple.antitoneOn_of_mul_id_antitoneOn {H : } {s σ : } (hs : 0 < s) (hH0 : tSet.Icc s σ, 0 H t) (hmono : AntitoneOn (fun (t : ) => H t * t) (Set.Icc s σ)) :

            The source hypothesis that t ↦ H(t)t is antitone, together with nonnegativity and positive coordinates, implies that H itself is antitone.

            The supported primes in [w,z), sorted and cast to real nodes.

            Equations
            Instances For

              The complete real node list: the genuine lower endpoint, all supported prime jumps, and the genuine upper endpoint.

              Equations
              Instances For
                theorem MathlibNt.SieveTheory.SwitchingPrinciple.suzukiFiniteNodeAtom_eq_primeAtom (S : BoundingSieve) (z y : ) (p : ) (hp : p S.prodPrimes.primeFactors) (hpz : p < z) (hpy : p < y) (hgap : qS.prodPrimes.primeFactors, p < qq < yFalse) :

                If y is the node immediately after a supported prime p, the successive node atom is exactly Suzuki's source prime atom. The hypotheses only express that no supported prime lies strictly between these two nodes.

                Suzuki finite-node Abel summation. Unlike a natural-interval version, its boundary terms are exactly g w and g z. The summands on the left are the successive local-ratio (prime-jump) atoms.

                The same specialization with the left side visibly written using Suzuki's successive local-ratio atoms.

                The real finite-node atom sum, with genuine endpoints w,z, is exactly Suzuki's finite prime sum for g x = H (log D / log x).

                theorem MathlibNt.SieveTheory.SwitchingPrinciple.transformedH_monotoneOn {D w z s σ : } {H : } (hD : 1 < D) (hw : 1 < w) (_hwz : w z) (hcoord : xSet.Icc w z, Real.log D / Real.log x Set.Icc s σ) (hH : tSet.Icc s σ, 0 H t) (hHt : AntitoneOn (fun (t : ) => H t * t) (Set.Icc s σ)) :
                MonotoneOn (fun (x : ) => H (Real.log D / Real.log x)) (Set.Icc w z)

                The Suzuki transform is increasing in the underlying variable.

                theorem MathlibNt.SieveTheory.SwitchingPrinciple.pairwise_imp_of_mem {α : Type u_1} {R T : ααProp} {l : List α} (hR : List.Pairwise R l) (hT : xl, yl, R x yT x y) :
                theorem MathlibNt.SieveTheory.SwitchingPrinciple.transformedH_suzukiFiniteNodes_increments_nonneg {D w z s σ : } {H : } (S : BoundingSieve) (hD : 1 < D) (hw : 1 < w) (hwz : w z) (hcoord : xSet.Icc w z, Real.log D / Real.log x Set.Icc s σ) (hH : tSet.Icc s σ, 0 H t) (hHt : AntitoneOn (fun (t : ) => H t * t) (Set.Icc s σ)) :
                List.Pairwise (fun (x y : ) => 0 H (Real.log D / Real.log y) - H (Real.log D / Real.log x)) (suzukiFiniteNodes S w z)

                Every forward finite-node variation increment g y - g x is nonnegative. In particular this holds for each adjacent pair used by the Abel variation sum.

                Exact finite-node decomposition of Suzuki's local Euler ratio into its log z / log x main ratio and the dimension-one error.

                Source-faithful exact main/error split for the Suzuki Lemma 8.6 prime sum. The upper endpoint error vanishes when log z ≠ 0.

                theorem MathlibNt.SieveTheory.SwitchingPrinciple.suzukiFiniteErrorVariation_le {S : BoundingSieve} {D w z s σ K : } {H : } (hD : 1 < D) (hw2 : 2 w) (hwz : w z) (hs : 0 < s) (hsσ : s σ) (hz : z = D ^ (1 / s)) (hw : w = D ^ (1 / σ)) (hH0 : tSet.Icc s σ, 0 H t) (hHt : AntitoneOn (fun (t : ) => H t * t) (Set.Icc s σ)) (hK : 0 K) (hlocal : HasDimensionOneLocalProductBound S K) :

                Suzuki Lemma 8.6's finite error-variation estimate. Only the one-sided local-product error bound is used; no absolute-value estimate for E is needed.

                theorem MathlibNt.SieveTheory.SwitchingPrinciple.suzukiLemmaEightSixDimensionOne {S : BoundingSieve} {D w z s σ K : } {H : } (hD : 1 < D) (hw2 : 2 w) (hs : 0 < s) (hsσ : s σ) (hz : z = D ^ (1 / s)) (hw : w = D ^ (1 / σ)) (hHcont : Continuous H) (hH0 : tSet.Icc s σ, 0 H t) (hHt : AntitoneOn (fun (t : ) => H t * t) (Set.Icc s σ)) (hK : 0 K) (hlocal : HasDimensionOneLocalProductBound S K) :
                suzukiLemmaEightSixPrimeSum S D w z H (1 / s * (t : ) in s..σ, H t) + 2 * K * H s / Real.log w

                Suzuki Lemma 8.6 in dimension one, with the source-faithful main term (1/s) ∫ₛ^σ H(t) dt. It follows from the generic local Euler-product bound; no prime-specific PNT or Mertens theorem is used.