Practical and Robust antenna downtilt Optimisation Approach in 5G Network

Optimizing base station (BS) configurations, particularly antenna tilts, is vital for cellular networks. Adjusting radiation patterns through tilts enhances cell coverage and mitigates interference. This paper proposes a practical approach to optimize tilt angles for a real-world non-stand-alone (NSA) 5G network, without costly options of adding new cell towers. Using a black-box surrogate model based on industry-standard 5G networks simulations, we employ a novel robust hybrid Bayesian optimization and non-dominated sorting genetic algorithm (BO-NSGA) framework to maximize signal-to-interference-plus-noise ratio (SiNR) and downlink reference signal received power (RSRP) coverage. Using a load balancing constraint besides robust optimisation enhances consistency and reliability of measures. Integrating demographic-based weighting for signal measures and Radio Access Technology (RAT) selection between 4G LTE and 5G new radio (NR) in NSA architecture further refines NSA solutions and yields practical solutions from suggested portfolios. Results demonstrate superior performance over existing 4G tilt settings, reducing RSRP under-coverage and boosting SiNR.