Prediction of Radio Resource Allocation in Addis Ababa’s LTE Network
No Thumbnail Available
Date
2025-11
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
The rapid growth of mobile data traffic in Addis Ababa places increasing pressure on Ethio Telecom’s LTE network and highlights weaknesses in existing radio resource management practices. This thesis develops and evaluates a predictive framework to improve downlink Physical Resource Block (PRB) allocation by forecasting PRB demand and enabling proactive resource management. Using real operational data collected from Ethio Telecom for a multi-month period, the study formulates PRB utilization prediction as a univariate time-series problem and implements Long Short-Term Memory (LSTM) neural networks trained on preprocessed hourly and weekly PRB usage traces. The thesis describes the end-to-end pipeline: data collection and cleaning, feature preparation, model training and hyperparameter tuning, and evaluation using standard forecasting metrics (RMSE, MAE, MAPE, and R²). Results indicate the proposed model produces accurate short-term PRB utilization forecasts and, when integrated with a reactive allocation policy, can reduce periods of congestion and improve throughput relative to static allocation strategies. The contributions include (1) a contextualized dataset and preprocessing approach for Addis Ababa’s LTE environment, (2) an LSTM-based forecasting model adapted for PRB utilization prediction, and (3) a practical framework for integrating forecasts into operational PRB allocation.
Description
Keywords
LTE, Physical Resource Block (PRB), predictive analytics, LSTM, radio resource allocation, time-series forecasting