Pseudorandom number generators in empirical tests

A reliable random number generator (RNG) is a basic and essential ingredient for any stochastic simulation. A single RNG is not always sufficient for Simulation. In many applications, several independent randorn numbers are required. Description and implementation of uniform combined multiple recursive generator is examined. We provide comparisons of the performance of some existing generators constructed as a grouping of two uniform combined multiple recursive generators (CMRG) and recommend the more promising one.