Graph Fick’s Neural Networks for Traffic Prediction and Resource Allocation in 6G Wireless Systems

Abstract

With previously unheard-of speed, capacity, and intelligence, 6G systems have the potential to completely transform connectivity in the rapidly changing wireless communication market. Efficient traffic prediction and subsequent resource allocation are key components of 6G network optimization. This paper presented a novel Graph Convolutional Networks with Energy valley based Fick’s Law Allocation (GCN-EVO-FLA) for traffic prediction and optimal resource allocation in 6g wireless system. The dataset was first pre-processed for the traffic prediction. Then the traffic can be predicted using the graph convolutional network and optimized the network parameters using Energy Valley optimizer. In addition, the optimal resource can be allocated using Fick’s Law algorithm (FLA). Finally, the performance of the proposed approach can be evaluated with the metrics RMSE, MAE, and Power consumption (PC) and compared with the existing methods. The proposed approach earned 97.32% of , 5.99 of MAE, 15.04 of RMSE and 873 (kWh) of power consumption. When compared to the existing method, the proposed method earned the best performance.