On Combining Pseudorandom Number Generators

Abstract

Let X = (X1,...,Xn) and Y = (Y1,...,Yn) be independent random vectors whose components take values in (O,1,...,m-1). Let r be the joint distribution of n independent random variables uniformly distributed on (O,1,...,m-1). We show that the distribution of Z = X + Y (mod m) is closer to r, in several metrics, than is either the distribution of X or of Y. The principle suggested by this result is that combining strings of pseudorandom numbers, generated by different generators, by addition mod m, will result in a string more random than any of the separate strings.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 15, 1976
Accession Number
ADA030693

Entities

People

  • Herbert Solomon
  • Mark O. Brown

Organizations

  • Stanford University

Tags

DTIC Thesaurus Topics

  • Algorithms
  • Applied Mathematics
  • Computer Programming
  • Computers
  • Data Science
  • Generators
  • Information Theory
  • Markov Chains
  • Mathematical Analysis
  • Monte Carlo Method
  • New York
  • Random Number Generators
  • Random Variables
  • Statistical Tests
  • Statistics
  • United States
  • United States Government

Fields of Study

  • Mathematics

Readers

  • Approximation Theory.
  • Parallel and Distributed Computing.