Skip to Main Content
In this paper, we study the performance of some of the key scheduling algorithms such as Proportional Fair Scheduler, Modified Longest Weighted Delay First and Exponential Rule. We also propose a new algorithm called EXPQW, a variant of the Exponential rule which assigns weights to the subscriber stations based on their queue length and waiting time thereby ensuring fairness and quality of service for non-real time applications. We also present three hierarchical schedulers which use a combination of the exponential rule for waiting time and queue-length and other scheduling rules. The algorithms have been implemented in OPNET Modeler 16. The scenarios studied are: the baseline configuration specified by the 802.16m Evaluation Methodology Document with up to 60 users in a service region and a saturated frame scenario with up to 120 users. The results indicate that EXPQW and the hierarchical schedulers have comparable throughput and fairness values with algorithms like Proportional Fair Scheduler and Modified Longest Weighted Delay First in moderately loaded and heavily loaded scenarios.