Construction of a family of finite maximal codes (Q1117704)

The author introduces a family \({\mathcal F}\) of finite maximal codes over a two-letters alphabet \(A=\{a,b\}\), defined as the set of finite codes C verifying: \[ C-1=p(A-1)S\quad \Leftrightarrow \quad A^*=SC^*P \] where P,S are finite subsets of \(A^*\) such that either P or S is contained in \(a^*\). It is shown that every finite maximal code with two b's is in \({\mathcal F}\). The main result of the paper consists on an algorithmical characterization of \({\mathcal F}\). Indeed, an algorithm based on the effective resolution of the inequality: \[ a^ M=\sum_{m\in M}(M,m)a^ m\leq a^{M+I+1}+a^ I \] where I is a given subset of \({\mathbb{N}}\) and M is a unknown subset of \({\mathbb{N}}\) is proposed in order to compute any code of \({\mathcal F}\). The paper ends with an application to the construction of all factorizations (T,R) of \({\mathbb{Z}}/n{\mathbb{Z}}\)- i.e., of subsets of \({\mathbb{N}}\) such that each element of \(\{\) 0,...,n-1\(\}\) can be written uniquely as a sum modulo n of an element of T and of an element of R - such that \(a^ T\) or \(a^ R\) is one factor of a polynomial factorization of \(P(n)=1+a+...+a^{n-1}\).