LTE Downlink Scheduling: A True Bayesian Estimate Approach

Abstract

Extended research has been made in exploring the possibilities of a better real-time oriented downlink scheduling. The reason for these possibilities is caused by a fast-paced growth demand for multimedia applications that are mainly developed for mobile devices, and requires a high-speed wireless transmission for its satisfaction. Repositioning mobile devices have been one of the challenges arising from that demand. Due to the growth of mobile device users, another challenge has also been found, which is the capability of wireless networks to handle multiple simultaneous users within a single cell network environment. Current downlink scheduling algorithm which can cope with this challenge, Most Largest Weighted Delay First (MLWDF), needs to be improvised to suits the demands. True Bayesian Estimate (TBE) is one of the Bayes Estimator models which is suitable for handling multivariate parameters. Three proposed TBE algorithms have been designed with each having a different key design and objective. TBE-Fair (TBE-F) has provided a fairer and less delay scheduling as compared to MLWDF while TBE-Delay (TBE-D) manages to have a higher throughput rate. TBE-Flow Delay (TBE-FD) is an overall scheduler that manages multivariate QoS to perform better for real-time scheduling. All the TBE’s algorithms have better performances than MLWDF in real-time traffic due to its main key design of real-time oriented scheduling which focuses more on video and VoIP flows.