New extension for MIXMAX pseudo-random number generation

Patrick Alken alken@colorado.edu
Sat Aug 12 23:29:00 GMT 2017


Dear GSL users,

  The developers of the MIXMAX pseudo-random number generators have
kindly made available their code with a GSL interface. I have added a
link to download the code on the GSL home page in the "Extensions" area.
I am far from an expert on random number generators, and rarely use them
in my own work, so I would greatly appreciate if any knowledgeable GSL
users could take a look at the code and provide any feedback. If people
think this would be valuable we can add it to GSL. Below is a
description of the new PRNGs from the authors

====

[The MIXMAX] project is devoted to the development of a new class of
pseudo-random number generators (PRNG) which are based on fine theorems
of ergodic theory,
a branch of mathematics dealing with Kolmogorov-Anosov  C-K-systems.
These systems are homogeneous hyperbolic systems with nonzero Kolmogorov
entropy. In short, they are maximally chaotic dynamical systems.

The generator has two main advantages: it is one of the fasts generators
on the market, and it generates high quality random sequences. The
articles and material of the
conferences devoted to the MIXMAX generators can be found on our webpage:

http://www.inp.demokritos.gr/~savvidy/mixmax.php

The MIXMAX generators have passed strong statistical tests, such as
Test-U01, the Dieharder, the PRACRAND and others. Importantly, their
mathematical properties are well-understood and accepted in the computer
science community as well. Our software is hosted on the High Energy
Physics software depositary HEPFORGE:

====

Some publications describing this work are:

On the Monte Carlo Simulation of Physical Systems J.Comput.Phys. 97, 566
(1991)

Matrix Generator of Pseudorandom Numbers J.Comput.Phys.97, 573 (1991)

The MIXMAX random number generator Comp. Phys. Commun. 196 (2015) 161

Thanks,
Patrick



More information about the Gsl-discuss mailing list