Sums of two squares and Heegner norms with few prime factors, XIV: the Buchstab fixed point of a vector sieve as an obstacle problem, causal computation, and variable dispersion levels

Part XIII sifted N − p by two semi-linear sieves in a vector sieve, iterated Buchstab's identity on the inert primes, and obtained N = Q_C′(u,v) + Q_C(x,y) with Q_C′(u,v) prime and Q_C(x,y) a P_r, with thresholds r₀(L) = 7, 10, 16 for the switching losses L = 1, 3/2, 3. This paper studies the fixed point of that recursion. In a suitable gauge it is an obstacle problem of Hamilton–Jacobi–Bellman type in the level X and the inert exponent a: either the rectangle (vector-sieve) bound is active, or a pure delay equation holds. The equation is causal in the level, so the fixed point is computed in a single sweep in O(n²) operations instead of O(mn³); it agrees with the iteration of Part XIII to machine precision and is about 2000 times faster. A half-level scaling law reduces the storage, and the free boundary is observed to degenerate. Results: with grids up to n = 12800 and a one-sided quadrature that errs on the safe side, the thresholds at the Bombieri–Vinogradov level 1/2 become r₀(L) = 7, 9, 15, and the D = 23 table of Part XIII becomes 9/8, 11/10, 17/16. Under a level of distribution of Lichtman's type for the switched sequence, uniform in the residue N (Hypothesis L, reduced to Lichtman's bilinear and divisor estimates), the thresholds are 6, 8, 14: every large even admissible N is p + x² + y² with Ω(x² + y²) ≤ 6. The same engine with a linear auxiliary sieve treats the problem of Nath and Xie (Acta Arith. 223 (2026), 51–82): primes p with p − 1 = m² + n² and p + 2 almost prime. Their numerical value Ω(p+2) ≤ 9 rests on an incorrect delay system for the linear sieve functions. With the correct functions, three constants of their switched term corrected and its levels chosen freely, their inequality gives Ω(p+2) ≤ 17, and the fixed point gives Ω(p+2) ≤ 15. The full audit is in the companion note https://github.com/Ruqing1963/nath-xie-sieve-audit. Code and data: https://github.com/Ruqing1963/two-squares-buchstab-fixed-point

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-14
DOI
https://doi.org/10.5281/zenodo.24231796
Primary Topic
Analytic Number Theory Research
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Sums of two squares and Heegner norms with few prime factors, XIV: the Buchstab fixed point of a vector sieve as an obstacle problem, causal computation, and variable dispersion levels

Ruqing Chen
Zenodo (CERN European Organization for Nuclear Research)
Analytic Number Theory Research
preprint

Sums of two squares and Heegner norms with few prime factors, XIV: the Buchstab fixed point of a vector sieve as an obstacle problem, causal computation, and variable dispersion levels

Ruqing Chen
preprint en

Abstract

Part XIII sifted N − p by two semi-linear sieves in a vector sieve, iterated Buchstab's identity on the inert primes, and obtained N = Q_C′(u,v) + Q_C(x,y) with Q_C′(u,v) prime and Q_C(x,y) a P_r, with thresholds r₀(L) = 7, 10, 16 for the switching losses L = 1, 3/2, 3. This paper studies the fixed point of that recursion. In a suitable gauge it is an obstacle problem of Hamilton–Jacobi–Bellman type in the level X and the inert exponent a: either the rectangle (vector-sieve) bound is active, or a pure delay equation holds. The equation is causal in the level, so the fixed point is computed in a single sweep in O(n²) operations instead of O(mn³); it agrees with the iteration of Part XIII to machine precision and is about 2000 times faster. A half-level scaling law reduces the storage, and the free boundary is observed to degenerate. Results: with grids up to n = 12800 and a one-sided quadrature that errs on the safe side, the thresholds at the Bombieri–Vinogradov level 1/2 become r₀(L) = 7, 9, 15, and the D = 23 table of Part XIII becomes 9/8, 11/10, 17/16. Under a level of distribution of Lichtman's type for the switched sequence, uniform in the residue N (Hypothesis L, reduced to Lichtman's bilinear and divisor estimates), the thresholds are 6, 8, 14: every large even admissible N is p + x² + y² with Ω(x² + y²) ≤ 6. The same engine with a linear auxiliary sieve treats the problem of Nath and Xie (Acta Arith. 223 (2026), 51–82): primes p with p − 1 = m² + n² and p + 2 almost prime. Their numerical value Ω(p+2) ≤ 9 rests on an incorrect delay system for the linear sieve functions. With the correct functions, three constants of their switched term corrected and its levels chosen freely, their inequality gives Ω(p+2) ≤ 17, and the fixed point gives Ω(p+2) ≤ 15. The full audit is in the companion note https://github.com/Ruqing1963/nath-xie-sieve-audit. Code and data: https://github.com/Ruqing1963/two-squares-buchstab-fixed-point

Zenodo (CERN European Organization for Nuclear Research)
Analytic Number Theory Research
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon
Sums of two squares and Heegner norms with few prime factors, XIV: the Buchstab fixed point of a vector sieve as an obstacle problem, causal computation, and variable dispersion levels — Ruqing Chen · Zenodo (CERN European Organization for Nuclear Research) (2026) | TGRS Research Map | TGRS