turkmath.org

A rank metric code is a code consisting of $n\times n$ matrices with the distance function $d(X, Y ) := rank(X -Y )$. Rank metric codes have close ties to subspace codes, which have important applications in network coding. Maximum rank distance (MRD) codes are rank metric codes $C$ meeting the Singleton-like bound $|C| = q^{n(n-d+1)}$, where $d$ is the minimum distance. Linear MRD-codes for each parameter were constructed by Gabidulin in 1985. The first non-trivial example of a non-linear MRD-code was recently given by Cossidente, Marino and Pavese for the case $n = 3, d = 2.$ In the case $n = d$, MRD-codes correspond to .nite (pre)semifields, that is, nonassociative division algebras. Semifields have received much attention in recent years, though their applications to coding theory have not been exploited to date. In this talk we will give an overview on semi.elds, introduce a new family of linear MRD-codes for each parameter, and use some of the theory of semi.elds to prove that these are inequivalent to the Gabidulin codes. We will also present some results on constant distance rank metric codes, which correspond to sets of matrices where the rank of the difference of two elements is fixed. We will pay particular attention to the case where all elements are symmetric or hermitian.