Evaluation of AI Models for Intelligent Network Slicing in OpenRAN: A Comprehensive Benchmarking Approach

The integration between Artificial Intelligence (AI) and Open Radio Access Network (OpenRAN) holds great promise for intelligent resource allocation within 5G networks. This study explores the design and implementation of radio access network applications (xApp and rApp) within OpenRAN, towards optimizing network slicing. It benchmarks the performance among several AI models - Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), Long Short-Term Memory networks (LSTM), K-Nearest Neighbors (KNN), Decision Trees (DT) and Support Vector Machines (SVM)- to identify the most efficient resource allocation within a dynamic 5G environment. Results indicate that Decision Tree (DT) model outperformed other models in term of training time, accuracy and precision measures as well as F1 score and Root Mean Squared Error (RMSE). The DT model displays powerful classification abilities by effectively and efficiently classifying network slices to enhance OpenRAN-based 5G networks' resource management processes. These findings illustrate DT's suitability as an intelligent network slicing solution which may prove advantageous in optimizing resource allocation in 5G networks.