Hybrid QoE-Based Joint Admission Control and Power Allocation

In this paper, we propose a learning-based Quality of Experience (QoE) model for voice services based on real-world data that models the Mean Opinion Score (MOS) in terms of the Received Signal Strength Indicator (RSSI). Unlike earlier studies that used an objective approach to model QoE, one key feature of our study is the use of a hybrid approach to accurately evaluate QoE by learning human behaviors. We will also apply our model to a power allocation problem and formulate an optimization problem to maximize the sum of the QoE of users while guaranteeing the minimum data rate for each user. We show that the proposed hybrid model outperforms the conventional objective model in terms of average MOS and outage probability. Furthermore, users are more satisfied with the QoE maximization problem compared to the conventional rate maximization problem. Also, due to the limited power available to meet the needs of all users, we will introduce a joint power allocation and admission control problem. In our proposed approach, BSs monitor the outage probability and MOS of the overall system for each connection request to determine the service's perceived quality level and then decide whether or not to accept a new connection. The findings show a trade-off between the number of admitted users and their satisfaction levels, giving operators significant insight in terms of resource utilization.