Documentation

MathlibNt.AnalyticNumberTheory.Vaughan.VaughanPrefixReduction

Finite Vaughan decomposition and primitive-character prefix reduction #

This module is a purely finite algebraic layer between Vaughan's identity and weighted_primitive_prefix_maximal. It makes no Bombieri--Vinogradov claim and introduces no analytic hypothesis.

For arbitrary cutoffs u,v, the exact all-n decomposition is

Λ n = TypeI(n;u,v) + TypeII(n;u,v) + Small(n;v).

Here Type I is vaughanFirst - vaughanMiddle, Type II is vaughanThird, and the last term retains the complete exceptional range n ≤ v. The final theorem consumes this equality inside every primitive-character prefix and reduces its weighted square maximum to three explicit coefficient energies. No cutoff is selected or hidden.

The two Type-I pieces of Vaughan's identity, kept together with their correct relative sign.

Equations
Instances For

    The genuinely bilinear (large-large) piece of Vaughan's identity.

    Equations
    Instances For

      The complete small-factor exception. Keeping this term makes the identity valid without a premise such as v < n.

      Equations
      Instances For

        The finite Vaughan identity, valid for every natural n and every pair of cutoffs. This internalizes Mathlib's μ * log = Λ API through the exact finite identity in AnalyticNumberTheory.Sieve.VaughanIdentity.

        A scalar coefficient sequence multiplied by the von Mangoldt function.

        Equations
        Instances For
          noncomputable def AnalyticNumberTheory.LargeSieve.vaughanTypeICoeff (b : ) (u v : ) (n : ) :

          The Type-I coefficient sequence at the literal cutoffs u,v.

          Equations
          Instances For
            noncomputable def AnalyticNumberTheory.LargeSieve.vaughanTypeIICoeff (b : ) (u v : ) (n : ) :

            The Type-II coefficient sequence at the literal cutoffs u,v.

            Equations
            Instances For
              noncomputable def AnalyticNumberTheory.LargeSieve.vaughanSmallCoeff (b : ) (v : ) (n : ) :

              The small-factor coefficient sequence at the literal cutoff v.

              Equations
              Instances For

                Pointwise complex-linear form of the exact finite Vaughan identity.

                theorem AnalyticNumberTheory.LargeSieve.primitiveCharacterPrefix_vaughan_eq (b : ) (y q u v : ) (χ : PrimitiveCharacter q) :
                nFinset.Icc 1 y, vaughanLambdaCoeff b n * χ n = nFinset.Icc 1 y, vaughanTypeICoeff b u v n * χ n + nFinset.Icc 1 y, vaughanTypeIICoeff b u v n * χ n + nFinset.Icc 1 y, vaughanSmallCoeff b v n * χ n

                Exact consumption by one primitive character and one finite prefix. This is the algebraic interface used before taking character/modulus sums or prefix maxima; no estimate has yet been applied.

                theorem AnalyticNumberTheory.LargeSieve.primitiveCharacterPrefixMaxSquare_le_three (b bI bII bSmall : ) (N q : ) (χ : PrimitiveCharacter q) (hdecomp : nFinset.Icc 1 N, b n = bI n + bII n + bSmall n) :

                A finite pointwise three-piece decomposition on [1,N] is consumed by the primitive-character prefix maximum, uniformly in the character and modulus.

                Exact algebraic reduction of every primitive-character prefix maximum of b Λ to Type I, Type II, and the retained small range. All four finite parameters N,Q,u,v remain explicit.

                theorem AnalyticNumberTheory.LargeSieve.weighted_vaughan_prefix_large_sieve_ledger (b : ) (N Q u v : ) (hQ : 0 < Q) :
                qFinset.Icc 1 Q, q / q.totient * χ : PrimitiveCharacter q, primitiveCharacterPrefixMaxSquare (vaughanLambdaCoeff b) 0 N q χ 3 * ↑(N.log2 + 1) ^ 2 * primitiveLargeSieveConstant N Q * (nFinset.Icc 1 N, vaughanTypeICoeff b u v n ^ 2 + nFinset.Icc 1 N, vaughanTypeIICoeff b u v n ^ 2 + nFinset.Icc 1 N, vaughanSmallCoeff b v n ^ 2)

                The concrete algebraic consumer of the proved prefix-maximal primitive large sieve. Closing Standard BV from here still requires genuine Type-I and Type-II energy estimates; this theorem does not disguise those estimates as a Prop premise or a BV conclusion.