This is the authors' abstract. We don't add key points for this paper.
We study the largest minimum distance d₃(n,k) of ternary Euclidean linear complementary dual (LCD) codes of fixed dimension k. For k = 4 we determine d₃(n,4) for every length n, settling the 19 residue classes modulo 40 left open by Araya, Harada and Saito (2021). For k = 5 we prove, with computer assistance, that d₃(n,5) ≤ ⌊(81n−195)/121⌋ for every n, and we determine d₃(n,5) for every n ≥ 140 outside the residue class 67 modulo 121, in which two consecutive values remain possible; Araya, Harada and Saito had determined the eventual value in 3 of the 121 classes. Building on their simplex-code reduction, we study the set of values that their integer 3^(k−1)n − (3^k − 1)d/2 takes when d is any lower bound on the minimum distance of an LCD [n,k] code; we call this integer the defect. Closure properties that come from the classical codes formed by the points outside a projective subspace reduce existence to a short list of explicit codes. Non-existence is studied through a parametrisation by excess distributions; for k = 4 we show that the excess distribution has a prescribed nonsingular Gram matrix modulo 3, and a rank and determinant argument then proves non-existence without any computer search. For k = 5 the non-existence results that need a computer rest on a solver-free enumeration and on integer programs with the Gram matrix in normal form. The integer programs behind the upper bound have certificates of infeasibility in the VIPR (Verifying Integer Programming Results) format, produced by an exact rational solver and checked independently, while the exact value in 16 residue classes rests on runs of the CP-SAT solver of OR-Tools without certificates. All codes, programs, certificates and result files are publicly available.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Promjiraprawat et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: