Random Number Generation from Right-Skewed, Symmetric, and Left-Skewed Distributions |
| |
Authors: | Eberhard O. Voit,& Lorelei H. Schwacke |
| |
Affiliation: | Department of Biometry and Epidemiology, Medical University of South Carolina, Charleston 29425, USA. |
| |
Abstract: | Monte Carlo simulations have become a mainstream technique for environmental and technical risk assessments. Because their results are dependent on the quality of the involved input distributions, it is important to identify distributions that are flexible enough to model all relevant data yet efficient enough to allow thousands of evaluations necessary in a typical simulation analysis. It has been shown in recent years that the S-distribution provides accurate representations for frequency data that are symmetric or skewed to either side. This flexibility makes the S-distribution an ideal candidate for Monte Carlo analyses. To use the distribution effectively, methods must be available for drawing S-distributed random numbers. Such a method is proposed here. It is shown that S-distributed random numbers can be efficiently generated from a simple algebraic formula whose coefficients are tabulated. The method is shown step by step and illustrated with a detailed example. (The tables are accessible in electronic form in the FTP parent directory at http:@www.musc.edu/voiteo/ftp/.) |
| |
Keywords: | Monte Carlo simulation random number generation S-distribution skewness |
|
|