CNN and Random Forest based ML approaches for UE Resource Optimization in QoS driven - Sliced 5G Networks

The growing demand for diverse services and applications in 5G networks necessitates the development of efficient resource optimization techniques. Network slicing, a promising approach to address this challenge, enables the creation of multiple virtual networks within a shared physical infrastructure, each tailored to specific service requirements. The research work explores the application of two machine learning algorithms, Convolutional Neural Networks (CNNs) and Random Forests, for optimizing User Equipment (UE) resource allocation in QoS-driven sliced 5G networks. The proposed CNN-based approach effectively captures the spatial and temporal dependencies in network data, while the Random Forest algorithm provides robust performance in handling high-dimensional and complex data. A comprehensive evaluation demonstrates the superior performance of the proposed approaches compared to traditional methods, achieving significant improvements in terms of throughput, latency, and resource utilization. This research also provides a simulation suite for a network consisting of base stations and clients that possible scenarios of 5G can fit into and make analysis of different concepts easy.