Rule discovery for pseudorandom number generator based on cellular automata

Since the behavior of cellular automata strongly depends on their rules, it is essential to find appropriate rules for a specific application. This paper concern with rule discovery in cellular automata pseudorandom number generators. In this paper, we propose a new criterion to find high quality rules for CA randomizers. We make use of the proposed criterion to find appropriate rules of radius two and to construct a CA randomizer based on one-dimensional non-uniform cellular automata. Moreover, this paper presents and proves some new observations about the relationship between behavior of a CA and its size.