Teste de hipóteses bayesiano para intensidade de tráfego em filas M/M/s utilizando o algoritmo amostragem-reamostragem por importância (SIR)

Abstract

The study of the queues aims to equate operating costs with the efficiency of the system and allow the rational sizing of the infrastructure, human resources, the necessary machinery and installations, aiming at a better performance. To this end, some parameters are defined that aim to systematize the study of queues, among which traffic intensity stands out. For a more consistent study of these systems, Bayesian inference can be used, which is based on the association of prior knowledge about a given system by a specialist in the area, called priori, to the empirical knowledge obtained from the analysis of data collected in loco, using, for that, the Bayes Theorem. This monograph aims to establish a sensitivity interval for traffic intensity in M/M/s queues using the importance-resampling-initiated algorithm (SIR) with an informative a priori distribution using a triangular distribution, making use of a program developed in the R software and analyzing a practical situation involving observations of queues in a supermarket.