Dynamic Uplink and Downlink Resource Allocation using Deep RI Learning Approach in 5G-HetNets

Abstract

To anticipate the continuously shifting traffic and channel state and to adjust the TDD config on the spot in a high-mobility environment, a novel strategy is required. In order to flexibly allocate radio resources in real time, we investigate the routing mechanism in the high - speed and heterogeneous network and present a unique intelligent TDD setup approach based on extensive relevance feedback. To change the TDD Up/Down-link ratio by weighing rewards, the concept suggests using vibrant Q-value iteration-based recurrent neural networks with an expertise replay memory mechanism. This is accomplished by using a deep neural network to extract the characteristics of the knowledge in a complex network. When compared to traditional TDD resource allocation algorithms, simulation results demonstrate that the suggested strategy significantly improves networking performance in terms of speed and packet loss rates.