PSEUDORANDOM NUMBER GENERATOR ON CELLULAR AUTOMATA

Abstract

Subject of Research. The paper presents an algorithm for pseudorandom number generationbased on properties of cellular automata. Cellular automata have high potential, high speed of calculations, especially at realization in parallel architecture. Method. In the presented algorithm pseudorandom numbers are generated by means of rules of transitions in cells of the cellular automaton depending on templates of the neighborhood and the output data of cells of "neighbors". Through several transitions at the choice of a generation technique the sequence of pseudorandom numbers turns out from zeroes and units. Main Results. The developed algorithm is tested on NIST-tests. The results of testing have shown that the algorithm makes the sequence with uniform distribution with probability of 99-100%. Comparison of the proposed algorithm with linearly congruent method, the main up-to-date method of generation of pseudorandom numbers, is carried out on NIST-tests. According to all tests the developed generator of pseudorandom numbers has shown the best results. The algorithm has the high speed, easy realization and also scaling possibility. Practical Relevance. The generator can be used in various applications, such as the theory of coding or lightweight cryptography. The cryptographic firmness is reached at tests by standard quality estimation techniques for the generator of pseudorandom numbers

Keywords

• cryptography
• pseudorandom number generator
• cellular automata
• homogenous structures
• NIST
• random numbers