New statistical method for generating sequences of pseudo-random numbers

Abstract

Abstract The paper describes the algorithm of the precision generator (hereinafter the P-generator) of random variables. Its feature is the ability to reproduce samples of large and small amounts of data with high accuracy knowing only the type of distribution. The created program module, which is based on the algorithm described in the work, is capable of generating a sample according to four distribution laws: normal, exponential, Laplace and Pareto. The proposed statistical method is based on calculating the number of points using priori known density distribution formula of the selected law. Further generation of the found number of points is implemented using the built into C# programming language random numbers generator. Computational experiments, which are given in the work, demonstrate the accuracy of the algorithm and also allow one to conduct a comparative analysis with existing alternatives (such popular mathematical and statistical packages as Mathcad and Excel).