Simulation of Several Round-Robin Variants in M/G/1 queues

The concept of processor sharing introduces a heuristics to prevent long jobs from blocking short jobs in queue. In practice, round-robin has to be used. In the literature, it is generally assumed that newly arrived jobs are put at the end of the queue. However, for highly varying job demands, such an approach may perform poorly. By favoring newly arrived jobs, which are more likely to be short, the performance of round-robin can be improved considerably. In this paper, we study the performance of several round-robin variants as a function of time slice values for different classes of service time distributions.