| Summary: | 碩士 === 國立臺灣科技大學 === 資訊工程系 === 103 === As the technology of mobile communication networks evolves, how to allocate limited resources efficiently is still a hot issue to be solved. Aiming at this issue, two scheduling algorithms will be proposed for the downlink of LTE/LTE-A in this thesis to greatly enhance the fairness while guaranteeing the quality of service (QoS) of different service flows.
As one knows, scheduling addresses the selection of service flows according to QoS, channel quality, and etc. In the literature, the existing scheduling algorithms for LTE/LTE-A, e.g., M-LWDF, M-EDF-PF, Log Rule, EXP/PF, and FLS, could meet the QoS requirements for real-time (RT) service flows, but almost no compensation is provided for the non-real-time (NRT) service flows in these algorithms. Therefore, RT service flows always get higher priorities than NRT service flows, leading to the starvation and lower fairness for NRT service flows. To address such an issue, this thesis targets to design two scheduling algorithms to not only guarantee QoS of RT service flows but also enhance fairness of NRT service flows. Examining performance metrics, including throughput, fairness index, delay, delay jitter, and packet loss ratio, the proposed scheduling algorithms are evaluated accordingly with comprehensive comparison to some closely related algorithms in the literature. As revealed by our simulation results, the proposed algorithms successfully achieve better fairness while guaranteeing QoS of different service flows as compared to the closely related algorithms and are highly recommended for LTE/LTE-A.
|
|---|
