Generalized Hamming weights of linear codes from cryptographic functions (Q6112192)

Cryptographic Boolean functions (a mapping from \(\mathbb{F}_{2^n}\) to \(\mathbb{F}_2\)) can be a useful tool for constructing linear codes, especially with few weights. This work shows a general formula for computing the generalized Hamming weights of linear codes from the defining sets from Bent and Semi-Bent functions, which generalizes several known results. The Maiorana-McFarland class \(\mathcal{M}\) is the set of all the Boolean functions: \(f(x,y) =\text{Tr}_m(x\pi(y)) + g(y),\) where \(\pi\) is any permutation over \(\mathbb{F}_{2^m}\) and \(g\) is any Boolean function on \(\mathbb{F}_{2^m}.\) For a balanced Boolean function \(g\) on \(\mathbb{F}_{2^m},\) the partial spread class corresponding to the Desarguesian spread and containing bent functions of the form \(f(x, y)=g(\frac{x}{y})\) for \(y\neq 0\) is the subclass denoted by \(\mathcal{PS}_{ap}.\) The class \(H\) consists of all Bent functions \(f(x,y)=\text{Tr}_m(x\psi(\frac{y}{x}))\) for \(x\neq 0,\) where \(\psi\) is a permutation on \(\mathbb{F}_{2^m}\) such that for every \(\beta\in\mathbb{F}_{2^m},\) the function \(\psi(x)+\beta x\) is two-to-one. The authors determine the weight hierarchies of linear codes with few weights (under several conditions) from three famous Bent function classes: Maiorana-MacFarland; \(\mathcal{PS}_{ap}\); and \(H\). In particular, in all three cases it is proved that when \(1\leq r\leq m\) the corresponding code is 2-weights code, and when \(m < r \leq 2m \) we have a 3-weights code. A similar study is also performed of two semi-bent function classes: \begin{itemize} \item Maiorana-MacFarland \item summation of \(\mathcal{PS}_{ap}\) and \(H.\) \end{itemize}