ABSTRACT
Human beings can operate the table-look-up method quite easily, but its implementation on a computer poses some intriguing problems. This chapter considers the analogue of the inversion method for continuous random variables. It shows that uniform random variables are the building-blocks for the simulation of any other random variable. Complicated algorithms utilizing composition and rejection methods, and sometimes requiring the storage of a large number of constants, are designed for use on computers that work to high precision. The chapter considers a general method suitable for discrete and continuous random variables. It presents a discussion of this method with an illustration from M. Abramowitz and I. A. Stegun. The chapter provides some further discussion of methods for simulating normal random variables. The simplest way for human beings to simulate random variables is to use tables of realizations of such random variables, such as those by H. Wold, providing normal random variables, and those by Barnett, which provide exponential random variables.
