In this paper, our attention is focused on the application of RDS on M/M/1 retrial queues for which an analytical solution exists and can therefore be used as a basis for comparison between RDS and SRS methods. This paper is concerned on the evaluation of performance measures, especially the number of customers in the orbit. A comparison of both MC methods, through the study of the mean number of customers in the orbit, is carried out practically using different efficiency statistical criteria showing that RDS outperforms SRS. To achieve such results, a simulator was designed and realized under Linux using the getRDS generator for the generation of all input RDS samples and the built in random number generator of the computing environment for the generation of all input SRS samples.
Variance reduction of the mean number of customers in the orbit of M/M/1 retrial queues using sampling methods
3 Views
1 Downloads