A Complete Guide to Gamma Variate Generation

Abstract
Considerable attention has recently been directed at developing simpler and faster algorithms for generating gamma random variates (with general, not necessarily integral, shape parameter α) on digital computers. This paper surveys the current state of the art, which includes fifteen gamma algorithms applicable for α ≥ 1 and six that are applicable for α < 1. These algorithms are compared according to the criteria of speed and simplicity. General random variate generation techniques are explained with reference to these gamma algorithms. Computer simulation experiments on DEC and CDC computers are reported. Guidelines for some specific applications are given.

This publication has 24 references indexed in Scilit: