Minimal linear codes constructed from functions (Q2154886)

This paper studies minimal linear codes arising from a general construction of linear codes from \(q\)-ary functions. A codeword \(c\) in a linear code \(\mathcal C\) is called minimal if \(c\) covers only the multiples \(ac\) for all \(a\in\mathbb{F}_q\). A code \(\mathcal C\) is minimal linear if every codeword in it is minimal. For \(k\in\mathbb{Z}, k\geq 2, m=k-1\) define a function \(f : \mathbb{F}_q^m\to \mathbb{F}_m^q\) and \(D=D_f=\left\{\text{d}_x = (f (x), x) | x \in\mathbb{F}_q^m \setminus\{0\}\right\}\) then \[ C_f= C(D_f)=\left\{c(u, v)=((u,v)\cdot d_x)_{x\in\mathbb{F}_q^m\setminus\{0\}}=(uf (x) + v \cdot x)\mid u\in\mathbb{F}_q , v\in\mathbb{F}_q^m\right\}. \] Assuming that \(f(x)\neq \omega x, \omega\in\mathbb{F}_q^m\) then \(\mathcal C\) is a \([q^m-1,m-1]_q\) linear code. The authors present new necessary and sufficient conditions for codewords by a general construction of linear codes from \(q\)-ary functions to be minimal. This is done by considering the three cases: (i) \(u\neq 0, v=0\); (ii) \(u\neq 0, v\neq 0\); and (iii) \(u=0, v\neq 0\). These new necessary and sufficient conditions does not require the knowledge of the weight distribution of the code and are thus better then previous. Using these necessary and sufficient conditions, three constructions of minimal linear codes that are generalizations of previously known constructions are shown. This is done by using much more flexible functions than previous results.