Efficient Algorithms for Computer Simulation of Gaussian Distribution
摘要
Abstract
The paper compares new efficient versions of the double-sided rejection algorithms with piecewise constant majorant and minorant and the modified ziggurat algorithm applied to computer modeling of the widely used Gaussian distribution. It is shown that these algorithms are significantly more cost-efficient than all known algorithms for modeling the Gaussian distribution (including the Box–Müller formulas, which are considered the most efficient).