Random Sampling from the Exponential Power Distribution

Abstract

Abstract Two simple and easily implemented algorithms are presented for obtaining random variables from the exponential power distribution with parameter α. Both algorithms are based on a generalization of Von Neumann's rejection technique. In the first algorithm, the first-stage sampling is from the double-exponential distribution, while the second algorithm uses the normal distribution. These two algorithms are applicable for all values of α, α ≥ 1 and α ≥ 2, respectively.