Author(s) :

Dahiru Sani Shuaibu, Shuaibu Alhassan Usman.

Volume/Issue :

Abstract :

Abstract-- Long term evolution (LTE) is one of the fastest growing technologies which support different type of applications . One of the key radio resources management mechanisms in the LTE network is the radio scheduler, which coordinate the access to radio resources, the decision of scheduling scheme play a major role in end to end LTE network performance and user Quality of service (QoS). Many resource scheduling scheme have been proposed and implemented in time past, among which is the modified largest weight delay first (MLWDF) algorithm which is suitable for real time and non-real time applications, this algorithm have been found not to support application sufficiently at overload network condition this is because it consider only head of line packet delay (HOL) as QoS parameter in making scheduling decisions which is not sufficient enough to improve overload network performance. This research evaluate some of the well-known algorithm in literature and make an improvement on the MLWDF by first adopting the mathematical model of MLWDF algorithm and then make an improvement on the adopted model by incorporating a scheduling priority ratio which is the ratio of (HOL) and delay deadline together with natural exponential term which help to grow the metrics for the users as their delay tolerance is approaching threshold thereby preventing packet loss and increase throughput in overload network condition. The proposed algorithm was simulated using LTE-Sim simulator and compared with other schedulers such as Flow bandwidth MLWDF for LTE downlink transmission (I-MLWDF), Improved MLWDF scheduler for LTE downlink transmission (Mod_MLWDF) and MLWDF using packet loss ratio, packet delay, throughput, fairness index and cell-spectral efficiency as performance metrics. The result shows that the proposed algorithm improved the video flow throughput at user index 30 by 38.33%, 59.35% and 60.45% and at user index 50 by 47.93%, 77.16% and 77.36% for VoIP flow it improve the throughput at user index 50 by 1.69% and 1.49%, for non-real time flow i.e, IMS signaling the proposed algorithm shows poor performance at user index 10, 30 and 50 by 66.91% ,88.19% and 90.71% respectively.

No. of Downloads :

4