Twister Generator of Random Normal Numbers by Box-Muller Model

Abstract

Twisting generators of the pseudorandom normal variables can use uniform random sequences as a basis. However, such technique could provide poor quality result in cases where the original sequences have insufficient uniformity or skipping of random values. This work offers a new approach for creating the random normal variables using the Box-Muller model as a basis together with the twisting generator of uniform planes. The simulation results confirm that the random variables obtained have a better approximation to normal Gaussian distribution. Moreover, combining this new approach with the tuning algorithm of basic twisting generation allows for a significantly increased the length of created sequences without using any additional random access memory of the computer.