QoE-Aware Scheduling for Sigmoid Optimization in Wireless Networks

This letter proposes a quality-of-experience (QoE)-aware scheduler that maximizes the average number of satisfied users. QoE is different from quality-of-service (QoS) in that QoE is determined by users' subjective satisfaction while QoS is objectively measured. Hence, users' satisfaction needs to be incorporated into the scheduler to better understand QoE maximization. In this regard we propose a simple but effective user participation scheme exploiting one-bit feedback that indicates user's satisfaction upon the end of service. Then, user-centric QoE function is derived, which serves our objective function. Assuming that the QoE function can be successfully modeled by a sigmoid function we solve the non-convex sigmoid optimization. We show that, interestingly, adding non-trivial fairness constraints for starvation prevention can convexify the problem, and we obtain the unique optimal solution. By comparing the case without user participation we show that user participation can substantially improve the total QoE, e.g., nearly 100% improvement of the low 10 percentile users and 40% improvement of the low 50 percentile users.