An overview on skew constacyclic codes and their subclass of LCD codes (Q825927)

Linear complementary dual (LCD) codes were introduced by Massey in 1994 [\textit{X. Yang} and \textit{J. L. Massey}, Discrete Math. 126, No. 1--3, 391--393 (1994; Zbl 0790.94022)] and proved that asymptotically good LCD codes exists [\textit{J. L. Massey}, Discrete Math. 106/107, 337--342 (1992; Zbl 0754.94009)]. From 2016 to 2018, some authors studied LCD codes [\textit{C. Carlet} et al., Des. Codes Cryptography 86, No. 11, 2605--2618 (2018; Zbl 1401.94199); \textit{C. Güneri} et al., Finite Fields Appl. 42, 67--80 (2016; Zbl 1364.94605); \textit{C. Li}, Des. Codes Cryptography 86, No. 10, 2261--2278 (2018; Zbl 1408.94982); \textit{C. Li} et al., IEEE Trans. Inf. Theory 63, No. 7, 4344--4356 (2017; Zbl 1370.94585); \textit{A. Sharma} and \textit{T. Kaur}, Cryptogr. Commun. 10, No. 3, 401--435 (2018; Zbl 1387.94124)], however few work is done on LCD codes in the noncommutative case, that is for LCD codes defined over skew polynomials rings \(\mathbb{F}_q[x;\theta]\). The paper in question deals with a first characterization of LCD skew constacyclic codes and some constructions of LCD skew cyclic and skew negacyclic codes over \(\mathbb{F}_{p^2}\). In Section 3, conditions for the equivalence between skew \(\lambda\)-constacyclic codes, skew cyclic codes and skew negacyclic codes are given. Let \(\theta\) be an automorphism of \(\mathbb{F}_q\). For an element \(\alpha \in \mathbb{F}_q\), the \(i\)-th norm is defined by \(N_i(\alpha) = \theta^{i-1}(\alpha) \cdots \theta(\alpha) \alpha\) and we have Theorem 3.1. If \(\mathbb{F}_q^*\) contains an element \(\alpha\) where \(\lambda=N_n(\alpha^{-1})\) then the skew \(\lambda\)-constacyclic codes of length \(n\) over \(\mathbb{F}_q\) are equivalent to the skew cyclic codes of length \(n\) over \(\mathbb{F}_q\). If \(\mathbb{F}_q^*\) contains an element \(\alpha\) where \(\lambda=-N_n(\alpha^{-1})\) then the skew \(\lambda\)-constacyclic codes of length \(n\) over \(\mathbb{F}_q\) are equivalent to the skew negacyclic codes of length \(n\) over \(\mathbb{F}_q\). In Section 4, the authors introduce the notion of LCD skew constacyclic codes and give some characterizations of their skew generator polynomials. In Theorem 4.1, necessary and sufficient conditions for skew \((\theta,\lambda)\)-constacyclic codes to be LCD codes (for both Euclidean and Hermitian duals) are given for \(\lambda^2=1\). Similar results are given in Theorem 4.3 for \((\theta,\lambda)\)-constacyclic codes in terms of the polynomials \(f_i\) (not necessarily irreducible) in the factorization of \(x^n -\lambda = f_1(x^\mu)^{p^s} \cdots f_r(x^\mu)^{p^s}\) where \(n\) is a multiple of the order \(\mu\) of the automorphism \(\theta\). In the final section, the authors focus on the construction (see Algorithm 4) and the enumeration (see Proposition 7) of LCD skew cyclic and negacyclic codes over \(\mathbb{F}_{p^2}\) where \(\theta\) is the Frobenius automorphism \(x \mapsto x^p\). When the skew generator polynomials of the codes are not divisible by any central polynomial, Proposition 7 gives nice explicit formulas for the number of Euclidean and Hermitian LCD \((\theta, \lambda)\)-codes of length \(2k\) and dimension \(k\). If \(p\) is odd, the Euclidean LCD skew cyclic codes of length \(2p^s\)and dimension \(p^s\) over \(\mathbb{F}_{p^2}\) are all Hermitian LCD codes. Interestingly, Remark 2 points out that, over \(\mathbb{F}_4\), the number of Euclidean self-dual \(\theta\)-cyclic \([2k,k]\)-codes, \(k=2^s\), is constant while the number of Euclidean LCD \(\theta\)-cyclic \([2k,k]\)-codes is \(2^k\), hence with exponentially growth in \(k\). The paper is well written and has several examples and tables to illustrate the results.