Documentation

MathlibNt.SieveTheory.Distribution.LiuPan.LiuPanConvolutionSourceCount

theorem MathlibNt.SieveTheory.LiuWeight.liuMainPanCoprimeIntervalSum_eq_count_sub_main (main : ) (N z y A₁ A₂ q l : ) :
liuMainPanCoprimeIntervalSum main N A₁ A₂ q l (liuWeight N z y) = liuCoprimeIntervalCount N z y A₁ A₂ q l - aFinset.Ioc A₁ A₂, if a.Coprime q then liuWeight N z y a * (main (N / a) / q.totient) else 0

Identification with the counting component of the actual production sum. The main function, its argument N/a in real arithmetic, and its normalization are left literally unchanged.

theorem MathlibNt.SieveTheory.LiuWeight.liuCoprimeIntervalCount_le_sum_beta (N z y A₁ A₂ q l : ) :
liuCoprimeIntervalCount N z y A₁ A₂ q l nliuAPCarrier N q l, liuBeta N z y n

Explicit domination by the untruncated beta sum, after whole-a regrouping.

theorem MathlibNt.SieveTheory.LiuWeight.liuSourceCoprimeIntervalCount_le (N A₁ A₂ q l : ) (hq : 0 < q) :
(∑ aFinset.Ioc A₁ A₂, if a.Coprime q then liuWeight N (liuSourceZ10 N) (liuSourceY3 N) a * (AnalyticNumberTheory.Sieve.primesInAPBelow N a q l) else 0) 3 * (N / q + 1)

Source-specialized counting estimate, uniform in arbitrary interval cuts, including reversed/empty intervals. No reduced-residue hypothesis is used.

The exact interval in the Pan--Ding--Wang consumer.

theorem MathlibNt.SieveTheory.LiuWeight.liuCoprimeIntervalCount_eq_sum_beta_of_support (N z y A₁ A₂ q l : ) (hl : l.Coprime q) (hsupp : ∀ (a : ), liuWeight N z y a 0a Finset.Ioc A₁ A₂) :
liuCoprimeIntervalCount N z y A₁ A₂ q l = nliuAPCarrier N q l, liuBeta N z y n

With genuine support containment and a reduced residue, neither retained filter removes an effective divisor. The support condition is explicit: arbitrary truncations admit only the domination theorem above.

Exact full-beta regrouping for the source interval whenever its known support-containment threshold holds. This is the only place reducedness is needed; none of the upper bounds require it.