Hybrid-boosted model with an approach inspired by a mixture of experts for 5G energy consumption

The advancement of 5G networks has brought significant innovations in terms of services and technologies but this has also imposed challenges related to energy consumption. Over 70% of energy consumption was projected to be attributed to Radio Access Networks (RANs), specifically Base Stations (BSs), with data centers and fiber transport contributing to a lesser extent. The objective of this study was to optimize the parameters of BSs and energy-saving methods, providing a deep understanding of how these elements influence energy consumption. This study introduces a hybrid-boosted ensemble model tailored for predicting energy utilization in 5G base stations. The methodology merges ridge regression for linear trend analysis, XGBoost to tackle non-linear intricacies, and a final refinement strategy built upon a mixture of experts. Preliminary results demonstrate the model's capability to adjust to new and diverse data scenarios, achieving up to a 31.94% improvement in 5G energy forecasting compared to traditional models. This research constitutes a pivotal initiative in confronting adaptability hurdles in 5G energy predictions, setting a foundation for subsequent inquiries in this essential domain.