Algorithms Synthesis of the Adaptive Logical Network on Basis of Universal Logical Elements

Abstract

An approach to the synthesis of adaptive structures represented by multi-level logic, Boolean network described as an acyclic graph, whose vertices are universal logical elements, is proposed. As a result of the synthesis of such structures are determined the types of logic functions of the vertices of a given learning sample of binary vectors, which allows using of this structure for the problem of classification of input vectors. Unlike known methods for the synthesis of multilevel logic, two approaches to the synthesis of such schemes proposed in this paper. The first way based on the description of a Boolean network represented by polynomials, the coefficients of the polynomial in this case are given by Hadamard matrix. The second way based on the description of Boolean network by Zhegalkin's polynomials considered in the article.