Cell Performance-Optimization Scheduling Algorithm Using Reinforcement Learning for LTE-Advanced Network

Abstract

Long Term Evolution Advanced (LTE-A) is one of the fastest growing technologies used for transmitting data in cellular networks. It provides its subscribers with enhanced service capabilities and enhanced network performance and this is carried through the smart deployment of new techniques and technologies. LTE-A is for improvement of the radio access part of cellular networks, for some time, it must co-exist with the 2G and 3G cellular networks, so interworking necessities, potential interference, resource management, etc. are an important issues. The Radio Resource Management (RRM) main role is to guarantee the efficient exploit of available radio resources using the available adaptation techniques, and to serve users based on their Quality of Service (QoS) parameters. In this paper, a novel dynamic neural Q-learning based scheduling algorithm is proposed for downlink transmission in LTE-A cellular network, aims to make a good trade-off between fairness and throughput. The proposed algorithm is based on the Q-learning technology and adoptable to variations in channel conditions. The key idea of our algorithm is how to intelligently choose the appropriate scheduling rule according to the predicted values of cell throughput and cell fairness index.