Proposta de um algoritmo para geração de números aleatórios em ambientes generalizados do canal sem fio com desvanecimento

Abstract

This thesis proposes a simple and highly eficient random sequence generator for statistical distributions used to model fading channels such as Nakagami-m, a-u, k-u, n-u, a-k-u e a-n-u. These generators are implemented via the Acceptance-Rejection technique using a majority function with estimated parameters, and achieve almost 100% of eficiency for all cases through mathematical transformations in generated samples with predefined parameters. Similar or better results can not be found in the literature. The algorithm effectiveness is demonstrated by the Kolmogorov-Smirnov test on the generated samples, proving its quality in relation to adherence with its respective theoretical distribution. One advantage of the proposed algorithm, plus its high effciency, comes from of there are not constraint parameters in the data generation, and how we treat of general distributions that encompass many of other particular cases, a single generator can be used for a wide range of diferent distributions through the correct parameters choice.