Proportional Fair Scheduling in Multi-Carrier Networks Using Channel Predictions

Proportional Fair Scheduling (PFS) is an important scheduling algorithm in wireless networks, aiming to utilize network resources efficiently while providing fairness among users. Recent progress in channel predictions makes it possible for the scheduler to take into account the achievable rates of users in the upcoming transmission intervals. In this paper we investigate the Predictive PFS in a multi-carrier system and show that the resulting optimization problem is a complex combinatorial problem in both time and frequency dimensions. We propose two suboptimal algorithms, and apply them to evaluate the effects of considering the predicted data rates on the performance of the network. Extensive simulations indicate improvements in both the achievable system throughput and access fairness among the users compared to the existing multichannel PFS method that does not use channel prediction.