Generate random variates using a newly introduced approximation to cumulative density of lower truncated normal distribution for simulation applications

Abstract

In this paper, the lower side truncated cumulative normal distribution is approximated by a simple function, the inverse of the function is derived, and random variates are explained how to be generated from the introduced inverse approximation. The introduced approximation is derived from Aludaat and Alodat's model of approximating cumulative normal distribution. The accuracy of the introduced function is investigated in term of maximum absolute error (i.e., 0.003944). This level of accuracy is possibly the best comparing all previous similar models to the best of the author's knowledge.