Computer Generation of Poisson, Binomial, and Hypergeometric Random Variates
Abstract
This research is focused on the development of exact, uniformly fast computer algorithms for generating random variates from the Poisson, binomial, and hypergeometric distributions. All four fundamental approaches to random variate generation, i.e., inverse transformation, composition, acceptance/rejection, and special properties, are used in the formulation of the new algorithms. The algorithms developed are implemented in FORTRAN and comparisons are made against comparable existing algorithms. All three algorithms developed dominate comparable existing algorithms in terms of execution speed. The one-sided approximations for the Poisson, binomial, and hypergeometric probabilities, which may be useful in a more general context, are given the appendices along with the computer codes.
Degree
Ph.D.
Subject Area
Industrial engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.