There is no cubic 8-circulant nut graph: completing the existence classification of cubic polycirculant nut graphs
Abstract
Bašić and Damnjanović, On cubic polycirculant nut graphs, arXiv:2411.16904v1 Conjecture 16 = Conjecture 1 of the published version (Comput. Appl. Math. 2025). Their Theorem 2 excludes l in {1,2,4,5} and constructs cubic l-circulant nut graphs for l in {3,6,7} and l >= 9; l = 8 was the only open case. Proof: a single kernel character on semiregular orbit coordinates gives a symmetric signed integral 8x8 quotient with nullity exactly one (injective lift) and nonzero diagonal cofactors; cubicity forces a local grammar; an inductive kernel-checked catalogue of 194 connected subcubic supports with 1,722 isomorphism certificates, 669 augmentations and tree switching reduce to 9,172 signed matrices, each excluded by an exact determinant (8,208) or zero principal cofactor (964) certificate. Valid for all orbit sizes m, without bounding graph order.
AI-assisted research preprint. The scoped main statement has a Lean certificate. Internal AI review is not external journal peer review; bounded prior-art searches do not certify worldwide priority. See the paper for exact qualifications and formalization limits.
Supplementary files
Comments
Submitted by Myeongjin Kang. Related MathDB problem 365943. Proof sources are attached as proof-sources.json (gzip-compressed, base64-encoded file contents); extract_sources.py reconstructs all archived files. Original archive SHA-256: e4710a538ed6ef60e6649fcb288dd873db52ce613eeddca564627f39c96ef402.