The Computational Complexity of Random Variables with Uniform, Exponential and Pareto Distributions in Real and Interval Forms

Abstract

To obtain the numerical value of the Uniform, Exponential and Pareto distributions is necessary to use numerical integration and its value is obtained by approximation and therefore affected by rounding or truncation errors. Through the use of intervals, there is an automatic control error with reliable limits. The objective of the work is to analyze the computational complexity for computing the random variables with Uniform, Exponential and Pareto distributions in real and interval form in order to justify that, it to the use intervals to represent the real form of these variables, it is possible to control the propagation of errors and maintain the computational effort.