Zusammenfassung (übersetzt):
The approximation of G/G/s models from Markov models M/M/s has been studied in the literature. The study of a queue model is detailed in the present article, using times of arrivals and time service distributed by Weibull whose estimation of parameters was performed with the Bayesian method Monte Carlo Markov chain, specifically the Gibbs sampler. The approximations of this model of waiting lines is evaluated by simulation. This methodology was applied to the case of delivery of refreshments to students of the University of Magdalena in Santa Marta, Colombia. The results show the utility and power to calculate indicators of a queue system when both, the arrival and attention times, are distributed as a Weibull.