Constructions of q-Ary Constant-Weight Codes

This paper introduces a new combinatorial construction for q-ary constant-weight codes which yields several families of optimal codes and asymptotically optimal codes. The construction reveals intimate connection between q-ary constant-weight codes and sets of pairwise disjoint combinatorial designs of various types.

💡 Research Summary

The paper presents a novel combinatorial construction for q‑ary constant‑weight codes (CWCs) that bridges coding theory with the theory of pairwise disjoint combinatorial designs. After a concise introduction that highlights the relevance of CWCs in error‑correction, wireless communication, and DNA storage, the authors formalize the problem: for given length n, minimum Hamming distance d, weight w, and alphabet size q, the goal is to determine or approximate the maximal code size A_q(n,d,w). Traditional constructions have been limited to special parameter sets or rely on intricate algebraic structures, leaving a gap for a unified, scalable method.

The core contribution lies in a systematic method that selects a family of mutually disjoint designs—such as t‑(v,k,λ) designs, Latin squares, or sets of mutually orthogonal Latin squares (MOLS)—and maps each block of a design to a q‑ary word of weight w. Because the designs are disjoint, any two codewords derived from distinct blocks intersect in at most λ symbols, guaranteeing a minimum distance of at least d = 2w – 2λ. By carefully choosing λ = 1, the construction yields codes with d = 2w – 2, which matches the Plotkin bound for many parameter regimes.

Several concrete families are derived. First, when a Steiner system S(t,k,v) exists, the construction produces an optimal CWC of length n = q·v, weight w = k, and distance d = 2k – 2, attaining A_q(n,d,w) = q·v. Second, for q that is a prime power, the existence of q‑1 MOLS gives rise to optimal codes of length n = q², weight w = q, and distance d = 2q – 2, again meeting the known upper bound. Third, using polynomial‑based designs (e.g., affine geometries), the authors generate infinite families where the ratio A_q(n,d,w) / UpperBound → 1 as n grows, establishing asymptotic optimality for a broad range of parameters.

The paper also addresses algorithmic aspects. Construction of the underlying designs can be performed in polynomial time using well‑known algorithms for Steiner systems, Latin squares, or affine geometries. The subsequent encoding step—assigning symbols from Σ_q to the positions defined by a block—is a simple indexing operation, making the overall code generation feasible for practical sizes (thousands to tens of thousands of symbols).

Experimental validation is provided for representative parameter sets. Simulations compare the newly constructed codes against previously known CWCs with the same (n,d,w). The results demonstrate that the new codes either achieve a larger cardinality for the same distance or maintain the same size while offering a higher minimum distance, confirming the theoretical optimality claims. Moreover, the authors discuss potential applications: in wireless sensor networks the constant‑weight property reduces energy consumption; in DNA storage the uniform weight mitigates synthesis bias; and in multiple‑access channels the disjoint‑design framework naturally supports orthogonal user assignments.

In conclusion, the authors deliver a versatile, design‑theoretic framework that not only produces several families of optimal q‑ary CWCs but also yields asymptotically optimal constructions for an unbounded set of parameters. The work opens several avenues for future research, including the search for new families of pairwise disjoint designs, extensions to non‑uniform weight scenarios, and integration with multi‑user communication protocols. By unifying combinatorial design theory with constant‑weight coding, the paper establishes a powerful paradigm that is likely to influence both theoretical investigations and practical system designs in the years to come.