This paper proposes a wireless network traffic prediction model based on long-term and short-term memory cyclic neural networks. Through simulation experiments, the throughput prediction of 5G wireless networks using different scheduling algorithms for many different types of services is studied. The results verify that the long short-term memory prediction model has acceptable prediction accuracy and algorithm training speed, meets the needs of wireless network traffic prediction, and has a good application prospect.

At present, the 5th-Generation (5G) wireless mobile communication standard has been released. 5G networks efficiently support enhanced mobile broadband traffic, ultra-reliable low-latency communication traffic, and massive machine-type communication. However, a major challenge for 5G networks is to achieve effective Radio Resource Management (RRM) strategies and scheduling algorithms to meet quality of service requirements. The Proportional Fair (PF) algorithm is widely used in the existing 5G scheduling technology. In the PF algorithm, RRM assigns a priority to each user which is served by gNodeB. The existing metrics of priority mainly focus on the flow rate. The purpose of this study is to explore how to improve the throughput of 5G networks and propose new scheduling schemes. In this study, the package delay of the data flow is included in the metrics of priority. The Vienna 5G System-Level (SL) simulator is a MATLAB-based SL simulation platform which is used to facilitate the research and development of 5G and beyond mobile communications. This paper presents a new scheduling algorithm based on the analysis of different scheduling schemes for radio resources using the Vienna 5G SL simulator.