Base Station Allocation For 6G Wireless Networks Using Wide Neural Network

Abstract

Terahertz (THz) transmission is a prominent technology in 6G mobile networks due to its enormous bandwidth and data transfer at fast speeds. in 6G wireless networks is crucial to managing massively increasing data rates and device connection for maximum performance and user experience optimum resource allocation is needed. It also allows dynamic network resource distribution for 6G's high device density and variable service requirements. Wide neural networks (WNN) can smooth network performance to address this issue. Paper proposes a WNN-based dynamic base station allocation method for 6G wireless networks. Training the WNN model with 14 6G parameters. Results show that the WNN strategy for dynamic decision-making in 6G networks works and might be used to other domains with comparable issues. With fewer fully connected layers, the wide neural network model performs better. Received validation accuracy is Interestingly, linear models without an activation function (None) perform as well as Tanh for single and two-layer topologies, with accuracy of 93% and 92%, respectively, and AUCs of 0.99. With three layers, accuracy decreases to 86%, still good.