Cyclotomic Digit-Reversal Thresholds: Explicit Slopes, Effective Periods, and Uniform Families

Primarily AI-generated textHuman understanding: no partsmath.NT — Number Theorymath.CO — Combinatorics

Contributed by Oleksiy Babanskyy ↗

Version 1 / Nov 18, 2026 / CC BY 4.0

Abstract

For a base-BB digit polynomial PP, let Kt(B)K_t(B) be the first length at which the represented integer and its reversal are divisible by the full integer Φt(B)\Phi_t(B), although Φt(X)\Phi_t(X) does not divide P(X)P(X). We study this threshold through a lattice of evaluations and the capacities imposed by digit positions. For every fixed t≥3t\geq3, we determine explicit linear slopes and construct an eventual affine description on residue classes, with computable periods, cutoffs and compressed minimum witnesses. No total-unimodularity assumption is needed. For every prime p≥5p\geq5, exact formulas in stated base ranges give K3p(10)=17p−1K_{3p}(10)=17p-1 and K6p(10)=18p+1K_{6p}(10)=18p+1. Ordered capacities explain their different correction periods, two and four. A deductive scalar classification also describes the minimum algebraic values for 6p6p and the action of reversal on them. Period transfers show that excluding palindromes leaves the first thresholds unchanged for both families. The same framework treats independent coefficient bounds and collisions. Exact results for orders 105,210105,210 require finite certificates beyond total unimodularity. The general construction is effective; minimal periods and practical efficiency are not claimed.

Author comments: 51 pages. Exact-arithmetic companion 1.1.0, including certificates and reproducibility instructions: https://github.com/aconsciousfractal/FCIG-Cyclotomic-Digit-Reversal-Thresholds/tree/28b00ab1a5d4ba64956ab82dafb02fc35ac40797. No independent specialist review or formal verification is claimed. Previously deposited as aiXiv 2609.03287 v1; the mathematical content and sources are unchanged.

Provenance statement

OpenAI Codex and GPT-6 Pro and Astra through ChatGPT substantially assisted mathematical exploration, proof development and checking, literature comparison, exact software and experiments, critical review, drafting and editing. This role is disclosed in the manuscript. The model names are reported by the author; exact builds are not recorded. Oleksiy Babanskyy directed the project and remains responsible for the article and its attributions. The author declares that no part of the contribution has been checked and understood by a human; AI reviews and computational replays are not represented as human understanding, independent specialist review or formal verification. The first-failure problem uses the full integer cyclotomic modulus and the stated digit alphabet. The general theorems concern each fixed order; the periods are effective but not asserted to be least or practically efficient as the order varies. Exact 3p and 6p formulas apply in their stated base domains. Threshold ranges, uniqueness ranges, normalized algebraic values and digit words are distinct. Orders 105 and 210 retain essential finite certificates; no finite replay replaces the uniform written proofs. General eventual quasi-polynomial behavior and witness selection are credited to Bogart-Goodrick-Woods, effective semilinear decomposition to Chistikov-Haase, and classical cyclotomic and total-unimodularity inputs to their cited sources. An author-side AI review on 28 September 2026 read all 21 TeX sources and appendices. Full verification passed normally and with Python optimization: 4 package tests and 40 mathematical tests per mode, and all F1-F8 components. A separately written script with no companion imports checked every nonconstant support and row for p=5,7,11,13,17,19, four base residues and two boundary prefixes: 15,977,952 support-boundary checks agree with the stated classification. It is a finite exposed check, not a proof for all primes. An isolated rebuild matches all 51 page texts and 51 page rasters of the deposited PDF. These are author-side computational and AI checks, not independent human peer review or proof-assistant verification. No exhaustive novelty or priority claim is made. This Hexagon transfer checks unchanged source identity, package closure, metadata and the compiled target PDF; it does not count as a new mathematical review. Prior deposit: https://aixiv.online/abs/2609.03287v1 Pinned source, exact certificates and reproduction instructions: https://github.com/aconsciousfractal/FCIG-Cyclotomic-Digit-Reversal-Thresholds/tree/28b00ab1a5d4ba64956ab82dafb02fc35ac40797

Tools used

OpenAI
Codex
OpenAI
GPT-6 Pro
OpenAI
Astra
Python Software Foundation
Python
SymPy
SymPyVersion 1.14.0

References

  1. J. Abbott, N. Mexis, Cyclotomic Factors and LRS-Degeneracy , arXiv:2403.08751v5 (https://arxiv.org/abs/2403.08751v5) (19 June 2026). Specific reference: §4.1.2--4.1.3 and Algorithm 3.arXiv
  2. T. Bogart, J. Goodrick, K. Woods, Parametric Presburger arithmetic: logic, combinatorics, and quasi-polynomial behavior , Discrete Analysis 2017:4. doi:10.19086/da.1254 (https://doi.org/10.19086/da.1254). Version consulted: arXiv:1608.08520v2 (https://arxiv.org/abs/1608.08520v2), Theorem 1.15 and Corollary 1.16.arXivDOI
  3. M. Beck, S. Hoşten, Cyclotomic Polytopes and Growth Series of Cyclotomic Lattices , Mathematical Research Letters 13 (2006), 607--622. Version consulted: arXiv:math/0508136v4 (https://arxiv.org/abs/math/0508136v4), Propositions 4--6 and Theorems 7, 13.arXiv
  4. J. L. Martin, V. Reiner, Cyclotomic and simplicial matroids , Israel Journal of Mathematics 150 (2005), 229--240. Version consulted: arXiv:math/0402206v1 (https://arxiv.org/abs/math/0402206v1), Theorem 1, Corollary 2, Remark 6, and §2--3.arXiv
  5. L. De Carli, M. Laporta, Divisibility criteria and coefficient formulas for cyclotomic polynomials , The Ramanujan Journal 70, article 44 (2026). doi:10.1007/s11139-026-01421-6 (https://doi.org/10.1007/s11139-026-01421-6). Comparison based on: arXiv:2508.02722v1 (https://arxiv.org/abs/2508.02722v1), Theorem 1.1.arXivDOI
  6. J. P. Steinberger, The lowest-degree polynomial with nonnegative coefficients divisible by the nn-th cyclotomic polynomial , The Electronic Journal of Combinatorics 19(4), P1 (2012). doi:10.37236/2755 (https://doi.org/10.37236/2755). The 18-page journal-distributed text, Theorem 1 and §2 (Lemma 1 and Proposition 1). The PDF retains provisional headers; bibliographic data follow the journal record.DOI
  7. A. Satyanarayana Reddy, The lowest degree 0,10,1-polynomial divisible by cyclotomic polynomial , arXiv:1106.1271v2 (https://arxiv.org/abs/1106.1271v2) (2011). §1--2: 0,10,1 multiples with at least three terms and minimal vanishing sums.arXiv
  8. M. J. Mossinghoff, Polynomials with Restricted Coefficients and Prescribed Noncyclotomic Factors , LMS Journal of Computation and Mathematics 6 (2003), 314--325. doi:10.1112/S1461157000000474 (https://doi.org/10.1112/S1461157000000474). §2.1--2.2, Algorithms 1--2.DOI
  9. D. Chistikov, C. Haase, The Taming of the Semi-Linear Set , ICALP 2016, LIPIcs 55, article 128. doi:10.4230/LIPIcs.ICALP.2016.128 (https://doi.org/10.4230/LIPIcs.ICALP.2016.128). Publisher version corrected on 26 October 2016, pp. 128:1--128:14, Proposition 5, pp. 128:5--128:6.DOI

Version history

  1. v1 · 51 pagesSubmitted by Oleksiy Babanskyy51 pages. Exact-arithmetic companion 1.1.0, including certificates and reproducibility instructions: https://github.com/aconsciousfractal/FCIG-Cyclotomic-Digit-Reversal-Thresholds/tree/28b00ab1a5d4ba64956ab82dafb02fc35ac40797. No independent specialist review or formal verification is claimed. Previously deposited as aiXiv 2609.03287 v1; the mathematical content and sources are unchanged.CurrentNov 17, 2026