AI-driven energy saving: traffic prediction and automated shutdown scheduling

Why AI Outperforms Static Schedules

Static time-based energy saving schedules (e.g., shutdown carrier 2 from midnight to 6 AM) typically save 5-10% because they use fixed rules that cannot adapt to variable traffic patterns. AI-driven approaches use machine learning models trained on historical traffic data (per-cell, per-hour, per-day-of-week) to predict low-traffic windows with high accuracy. When the model predicts the next 30 minutes will have <15% PRB utilization, it proactively schedules carrier shutdown or MIMO layer reduction. AI approaches achieve 15-30% energy savings because they adapt to events (holidays, stadium…