Pseudo-Random Number Generators

The first option here is not easy. Some pieces of data of this kind exist, for example from the decay of radioactive elements, but generally generation is slow and you cannot get the quantity of random numbers you require for modern applications. The second option here is what I’ll talk about today. Algorithms of this type are called “PseudoRandom Number Generators” (PRNGs). Many bad algorithms exist, in the sense that they generate numbers which can fairly readily be distinguished as having a pattern, and not being from a truly random process. I will discuss some of these issues and also outline the algorithms for some famous PRNGs, both good and bad. I’ll focus on PRNGs for generating pseudo-random numbers from the uniform distrubution on the interval [0, 1]. It is fairly easy to transform these for a different distribution if required.