LAUSR

To achieve higher performance of allocation resources in the densely distributed scene of 5G networks, resources can be allocated considering the communication traffic of different areas. For this requirement, a new allocation algorithm based on GCN-LSTM (Graph Convolution Network - Long Short-Term Memory) is proposed, which can provide an effective strategy for resource allocation. In this method, the historical 5G traffic data is divided into three time periods, and the spatial features of each period can be obtained through GCN. LSTM model is used to extract the temporal features. The combination of spatial and temporal features can achieve a high-precision prediction of 5G traffic data. Eventually, the 5G traffic prediction results of GCN-LSTM are used as the strategy basis for resource allocation. The simulation result demonstrates that the proposed algorithm can effectively improve the system throughput and spectral efficiency.