Dynamic Bandwidth Allocation in 5G Networks Using Multilayer Perceptron (MLP)

This paper presents a Multilayer Perceptron (MLP)-based model for dynamic bandwidth allocation in 5G network slicing. Traditional approaches to bandwidth allocation often suffer from rigidity and poor adaptability, particularly in handling diverse Quality of Service (QoS) requirements of heterogeneous services such as eMBB, URLLC, and mMTC. These limitations result in inefficient resource utilization and suboptimal user experience. To address this, an MLP model that dynamically classifies and allocates bandwidth with improved accuracy and responsiveness was developed. The model achieved 99.82% accuracy, 99.56% precision, 99.78% recall, and 0.03238 validation loss, demonstrating its ability to classify service types with exceptional precision. Compared to existing methods, the MLP model significantly outperforms prior works in classification performance and adaptability, making it a robust candidate for intelligent network slicing in next-generation networks.