Simulating polarization by random context filters in networks of evolutionary processors

Abstract

Networks of evolutionary processors (NEP for short) form a class of models within the new computational paradigms inspired by biological phenomena. They are known to be theoretically capable of solving intractable problems. So far, there are two main categories that differ from each other by the nature of filtering process controlling the communication step: random-context clauses or polarization. Several studies have proven that both of them are computationally complete through efficient simulations of universal computational models such as Turing machines and 2-tag systems. Nevertheless, the indirect conversion between the two network variants results in an exponential increase of the computational complexity. In this paper, we suggest a direct simulation of polarized NEP through NEP with random-context filters which incurs in lower complexity costs.