Semi-Persistent Queue-Assisted Carrier Aggregation for Non-GBR Traffic

User demand for bandwidth is ever increasing with time. Also, active user density becomes a limiting factor in cellular network even in 5G and beyond. Carrier aggregation (CA) technique was introduced in 3GPP Release 10 to circumvent this issue by providing more bandwidth through additional carrier. However, due to dynamic change in user demand & density, fixed carrier allocation scheme of CA does not guarantee better bandwidth always. In this paper, a novel scheme for allowing a user to request for additional carrier explicitly, is proposed. An optimization formulation for semi-persistent scheduling of users, considering user requests, has been constructed. Due to NP-Hardness of the formulated optimization problem, a$M/M/s/K$queue based user scheduling technique has been developed. The proposed scheduling technique decides about admissibility of the user request for carrier aggregation in a component carrier during a time-window. An admission control algorithm, based on queue results, has been presented and analyzed. A detailed simulation study is conducted to assess performance of the proposed scheduling technique over Proportional Fair (PF), Round Robin (RR) & Reinforcement Learning (RL) based schedulers.