High Quality Random Number Generator using Programmable Cellular Automata

Abstract

Random Numbers have applications in a number of computer sciences domains specifically in computer security algorithms they play a vital role. One of the techniques to generate a random number is to follow an algorithm which tries to generate random numbers as close as a true random number generator source, this algorithm based random number generation is known as Pseudo random Number Generation (PRNG). In this research we studied existing techniques of Cellular Automata (CA) based Pseudo Random generator (PRNG) and proposed 5 different variations in the existing algorithm, modified the CA rules as per requirement for the proposed techniques i.e. 3D Neuman and Moore neighborhood, 2D to 3D jumping Neuman and 2D Neuman and Moore jumping neighborhoods. In all the cases Neuman neighborhood means non diagonal adjacent neighbors while Moore neighborhood means diagonally adjacent neighbors. CA rules define which neighbors participate in the current cycle for random number generation. We implemented the simulation code in C++ and tested the outputs in internationally accepted standard for Random Numbers “DIEHARD”, the analysis of the results are made as per recommendations of ANSI standards and it is discussed which of the above techniques can be declared to be best in terms of quality of randomness in random numbers generation.