Möbius harmonic factor and centered progression-count bilinear sums
DefinitionGRH_MoebiusBilinearanalytic-number-theoryarithmetic-progressionsmoebius-function
For a positive modulus , let count satisfying , and put
The centered bilinear quantity subtracts the progression density for invertible :
These finite definitions separate the principal harmonic main term from cancellation in the arithmetic-progression rounding errors. All cutoffs are integer floors; empty sums at are zero.
Definition code
import Definitions.Def_GRH_SquarefreeBilinear
noncomputable section
open scoped BigOperators
namespace GRH
def moebiusProgressionCount (q : ℕ) [NeZero q] (a : ZMod q) (N d e : ℕ) : ℝ :=
(((Finset.Ioc 0 (N / (d * e))).filter
(fun r => ((d * e * r : ℕ) : ZMod q) = a)).card : ℝ)
def reducedMoebiusHarmonic (q : ℕ) [NeZero q] (U : ℕ) : ℝ :=
∑ d ∈ Finset.Ioc 0 U, if IsUnit (d : ZMod q) then
(ArithmeticFunction.moebius d : ℝ) / (d : ℝ) else 0
def centeredMoebiusBilinear (q : ℕ) [NeZero q] (a : ZMod q) (N : ℕ) : ℝ :=
∑ d ∈ Finset.Ioc 0 N.sqrt, (ArithmeticFunction.moebius d : ℝ) *
∑ e ∈ Finset.Ioc 0 N.sqrt, (ArithmeticFunction.moebius e : ℝ) *
(moebiusProgressionCount q a N d e -
if IsUnit (d : ZMod q) ∧ IsUnit (e : ZMod q) then
(1 / (q : ℝ)) * (N : ℝ) / ((d : ℝ) * (e : ℝ)) else 0)
end GRH
Source
Adapted finite notation for Prove2Me theorem GRH.moebius_reduced_progression_bilinear_power_bound, https://prove2.me/theorems/1e39fbca-7acc-40a2-baab-ab807852220a. The bilinear truncation is Helfgott–Thompson, Summing μ(n): a faster elementary algorithm, §2, equations (2.1)–(2.2), https://doi.org/10.1007/s40993-022-00408-8. Centering subtracts the elementary density 1/q in a reduced arithmetic progression; no analytic estimate is included.