Random Number Generators

Abstract

In contemporary computation there is an almost unquenchable thirst for random numbers. One particularly intemperate class of customers is comprised of the diverse Monte Carlo methods.¹ Or one may want to study a problem (or control a process) that depends on several parameters which one doesn’t know how to choose. In such cases random choices are often preferred, or simply convenient. Finally, in system analysis (including biological systems) random “noise” is often a preferred test signal. And, of course, random numbers are useful — to say the least — in cryptography.

“Anyone who considers arithmetical methods of producing random digits is, of course, in a state of sin.”

— John von Neumann