Vaughan's identity:
ProvedTaoFivePrimes.vaughan_identityWork in the ring of real-valued arithmetic functions under Dirichlet convolution, writing for the convolution, for the constant-one function, for the Möbius function, for the von Mangoldt function and for . For real cutoffs let and denote the truncations of an arithmetic function to arguments at most , respectively greater than . Then
This is Vaughan's identity in the exact form used to open the minor-arc analysis. Read pointwise at an integer and summed against a test function, the first two terms become Type I sums, in which one variable runs over a short range and the other carries a smooth weight; the third is the genuinely bilinear Type II sum handled by the large sieve; and the fourth is supported on and is negligible for the test functions used, which vanish there.
Formalization Note The truncations and the Dirichlet convolution are those of the ambient library; the constant-one function and the Möbius function are taken with real values so that all four terms live in the same ring.
import Mathlib import Definitions.Def_TaoFivePrimes_VaughanTruncation
theorem TaoFivePrimes.vaughan_identity (U V : ℝ) :
TaoFivePrimes.truncLe U TaoFivePrimes.moebiusR * ArithmeticFunction.log
- TaoFivePrimes.truncLe U TaoFivePrimes.moebiusR *
TaoFivePrimes.truncLe V ArithmeticFunction.vonMangoldt * TaoFivePrimes.zetaR
+ TaoFivePrimes.truncGt U TaoFivePrimes.moebiusR *
TaoFivePrimes.truncGt V ArithmeticFunction.vonMangoldt * TaoFivePrimes.zetaR
+ TaoFivePrimes.truncLe V ArithmeticFunction.vonMangoldt
= ArithmeticFunction.vonMangoldt := by sorry