Improving 4G Long Term Evolution Network Performance using Fairness-Aware and Adaptive Downlink Scheduling Algorithm

Abstract

Mobile network users desire uninterrupted communication, regardless of their mobility and distance, irrespective of the application type. As a result, both users and service providers place great value on Quality of Service (QoS). However, delivering such seamless connectivity is challenged by the limitations of wireless networks in supporting numerous users with varying service demands simultaneously. Consequently, ensuring fairness in resource allocation across different traffic classes and among users within the same class remains a key challenge for radio resource management (RRM) in Long Term Evolution (LTE) networks. The dynamic and unpredictable nature of wireless environments impaired by user mobility necessitates robust mechanisms for channel quality indicator (CQI) estimation. This paper introduces a new scheduling scheme, Fairness-Aware and Adaptive Downlink (FAAD) algorithm, designed to enhance the performance of 4G LTE networks. FAAD incorporates an Extended Kalman Filter (EKF)-oriented channel quality estimation and Unfairness Aversion Metric in computing scheduling metrics. The LTE-Sim software was used to assess the performance of the FAAD scheme against two existing schedulers: the Downlink Scheduling Algorithm with imperfect CQI (DSA) and the Enhanced Exponential Rule (EnEXP), in terms of packet delay, throughput, packet loss rate (PLR), fairness index, and jitter. Simulation results and statistical analysis, using Friedman's and Nemenyi's post-hoc tests, demonstrate that FAAD marginally outperforms the two existing schemes, exhibiting superior or competitive performance with respect to packet delay, PLR, and fairness, particularly under Voice over Internet Protocol traffic. The developed algorithm proves to be a viable option for enhancing LTE network performance, especially in scenarios with mixed traffic types