CDF-Based Scheduling Algorithm for Proportional Throughput Fairness

Abstract

Cumulative distribution function (CDF)-based scheduling (CS) with resource-based fairness is well-known as an efficient algorithm that satisfies fair resource sharing among users while accommodating exploitation of multiuser diversity. In this letter, we extend CS to satisfy proportional throughput fairness. In particular, the proposed CS with proportional throughput fairness (CSPT) exploits the properties of CS to intelligently calculate the channel access ratio (CAR) for each user that satisfies proportional throughput fairness. Then, user selection is performed under CS with the obtained CARs. Moreover, the throughput upper-bound of schedulers satisfying proportional throughput fairness is analyzed and, through asymptotic analysis and simulations for various scenarios, we show that the performance of our proposed CSPT is close to this upper-bound.