Sub-0.80 Asymptotic Upper Bounds and Analytic p-Adic OEIS Extensions for Erdős Problem #301

Let f(N) = A390394(N) denote the maximum size of a subset A ⊆ {1, ..., N} containing no distinct integers a, b₁, ..., bₖ (k ≥ 2) satisfying 1/a = 1/b₁ + ... + 1/bₖ (Erdős Problem #301). Erdős and Graham conjectured that f(N) = (1/2 + o(1))N. Building on the disjoint divisor-box framework of Wouter van Doorn (M = 12 → 25/28), Xinjun Wang (May 2026, M = 720 → 667/806), and Donald Della Pietra (July 30, 2026, who first broke below 0.80 with M = 2520 → 839/1040 and M = 5040 → 15437/19344), this preprint proves a sound summation-by-parts prefix-truncation inequality and certifies tighter exact rational asymptotic upper bounds at M = 10080 using two independent, soundness-hardened C99 exact branch-and-bound engines: • M = 5040 (complete, 60 divisors; independent reproduction of Della Pietra): lim sup f(N)/N ≤ 15437/19344 ≈ 0.798025 • M = 10080 (sound prefix j ≤ 57, d ≤ 504): lim sup f(N)/N ≤ 3481/4368 ≈ 0.796932 • M = 10080 (dual-engine verified sound prefix j ≤ 63, d ≤ 1008): lim sup f(N)/N ≤ 2233/2808 ≈ 0.795228 (improves 15437/19344 by 0.280 percentage points) Second, by an elementary zero-search p-adic non-cancellability argument (primes p = 73 and p = 37), we prove that adding 73 and 74 to {1, ..., 72} introduces no forbidden unit-fraction relations of any length k ≥ 2. Combined with William Alexander Carney's OEIS-approved value A390394(72) = 53 (independently re-confirmed from scratch), this establishes A390394(73) = 54 and A390394(74) = 55 (sharp, as 1/15 = 1/30 + 1/50 + 1/75). Included Files & Offline Reproduction • erdos301_arxiv.pdf: Compiled preprint (CC-BY-4.0). • erdos301_arxiv_upload.tar.gz: LaTeX source (erdos301_arxiv.tex) plus self-contained ancillary verification code (anc/erdos301_exact.c, anc/erdos301_complement.c, anc/verify_erdos301.py, anc/REPRODUCE.txt; MIT License). Unpack and run python3 anc/verify_erdos301.py (Python 3.11+ stdlib only) to verify all rational bounds, witnesses, and the p-adic proof, or compile both C99 engines with clang -O3 -Wall -Wextra -Werror -lm. Transparency Note: Computational exploration, C99 implementation, and manuscript drafting were assisted by AI agents; all proofs, rational arithmetic checks, witnesses, and soundness invariants were deterministically verified across two independent branch-and-bound engines.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-21
DOI
https://doi.org/10.5281/zenodo.24413617
Primary Topic
Analytic Number Theory Research
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Sub-0.80 Asymptotic Upper Bounds and Analytic p-Adic OEIS Extensions for Erdős Problem #301

Chimdindu Denalex Orakwue
Zenodo (CERN European Organization for Nuclear Research)
Analytic Number Theory Research
preprint

Sub-0.80 Asymptotic Upper Bounds and Analytic p-Adic OEIS Extensions for Erdős Problem #301

Chimdindu Denalex Orakwue
preprint en

Abstract

Let f(N) = A390394(N) denote the maximum size of a subset A ⊆ {1, ..., N} containing no distinct integers a, b₁, ..., bₖ (k ≥ 2) satisfying 1/a = 1/b₁ + ... + 1/bₖ (Erdős Problem #301). Erdős and Graham conjectured that f(N) = (1/2 + o(1))N. Building on the disjoint divisor-box framework of Wouter van Doorn (M = 12 → 25/28), Xinjun Wang (May 2026, M = 720 → 667/806), and Donald Della Pietra (July 30, 2026, who first broke below 0.80 with M = 2520 → 839/1040 and M = 5040 → 15437/19344), this preprint proves a sound summation-by-parts prefix-truncation inequality and certifies tighter exact rational asymptotic upper bounds at M = 10080 using two independent, soundness-hardened C99 exact branch-and-bound engines: • M = 5040 (complete, 60 divisors; independent reproduction of Della Pietra): lim sup f(N)/N ≤ 15437/19344 ≈ 0.798025 • M = 10080 (sound prefix j ≤ 57, d ≤ 504): lim sup f(N)/N ≤ 3481/4368 ≈ 0.796932 • M = 10080 (dual-engine verified sound prefix j ≤ 63, d ≤ 1008): lim sup f(N)/N ≤ 2233/2808 ≈ 0.795228 (improves 15437/19344 by 0.280 percentage points) Second, by an elementary zero-search p-adic non-cancellability argument (primes p = 73 and p = 37), we prove that adding 73 and 74 to {1, ..., 72} introduces no forbidden unit-fraction relations of any length k ≥ 2. Combined with William Alexander Carney's OEIS-approved value A390394(72) = 53 (independently re-confirmed from scratch), this establishes A390394(73) = 54 and A390394(74) = 55 (sharp, as 1/15 = 1/30 + 1/50 + 1/75). Included Files & Offline Reproduction • erdos301_arxiv.pdf: Compiled preprint (CC-BY-4.0). • erdos301_arxiv_upload.tar.gz: LaTeX source (erdos301_arxiv.tex) plus self-contained ancillary verification code (anc/erdos301_exact.c, anc/erdos301_complement.c, anc/verify_erdos301.py, anc/REPRODUCE.txt; MIT License). Unpack and run python3 anc/verify_erdos301.py (Python 3.11+ stdlib only) to verify all rational bounds, witnesses, and the p-adic proof, or compile both C99 engines with clang -O3 -Wall -Wextra -Werror -lm. Transparency Note: Computational exploration, C99 implementation, and manuscript drafting were assisted by AI agents; all proofs, rational arithmetic checks, witnesses, and soundness invariants were deterministically verified across two independent branch-and-bound engines.

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
Sub-0.80 Asymptotic Upper Bounds and Analytic p-Adic OEIS Extensions for Erdős Problem #301 — Chimdindu Denalex Orakwue · Zenodo (CERN European Organization for Nuclear Research) (2026) | TGRS Research Map | TGRS