A new construction for \(q\)-ary constant weight codes (Q6156927)

A \(q\)-ary constant weight \((n, d, w)\) code is a subset of \(\mathbb{F}_q^n\) with minimum Hamming distance \(d\) such that for every codeword we have \(\mathrm{wt}(c)=w.\) Such code \(C\) is said to be optimal if \(|C|=A_q(n,d,w)\) -- the maximum size of subset for which \(q\)-ary constant weight \((n, d, w)\) code exists. In this paper, the authors present a new construction for \(q\)-ary constant weight codes. This is achieved by following a two-step process: for any given non-empty finite sets \(T,\) \(D\) and \(S,\) let \(\Phi\) be a map from \(D\times T\) to \(S\). Choose some disjoint sets \(L_0(t),\ldots, L_{q-1}(t)\) with their union containing \(\Im_t(\Phi)\) then a \(q\)-ary code \(C_\Phi(D,T)\) with length \(|D|\) can be defined. Defining \(N_{L_i(t)}(t) = \{d \in D\mid \Im_t(\Phi)(d, t) \in L_i(t)\}\) for all \(t\in T\) it's proved that \(C_\Phi(D,T)\) is a constant weight code if and only if the value of \(|N_{L_0(t)}(t)|\) is independent of the choice of \(t.\) Furthermore, several special classes of \(q\)-ary constant weight codes are obtained. This is done using the trace map over finite fields. Especially for \(q=2 \) some of the obtained codes are proven to be optimal, and for a prime \(p\) and \(r, m\in \mathbb{Z}_{+}\) with \(r\vert m,\) it's shown that: \begin{itemize} \item \(A_2(p^{m-1}, 2(p^r-1)p^{m-2r},p^{m-r}-1)=\frac{p^m-1}{p^r-1},\) when \(2r\leq m;\) \item \(A_2(p^{m-r}, 2(p^r-1)p^{m-3r},p^{m-2r})=\frac{p^m-p^r}{p^r-1},\) when \(3r\leq m.\) \end{itemize}