Optimizing Nominal Current Output in the Elevated Runway Edge Lighting System at Subang Airport through Machine Learning Analysis

Aeronautical ground lighting systems are crucial for ensuring airport safety and operational efficiency. They provide essential visual guidance for pilots and ground vehicles, significantly minimizing the risk of runway incursions. However, improper usage of these lighting systems can lead to excessive power consumption, increase in electricity costs, and carbon emissions. This study conducts a thorough analysis of how meteorological parameters impact the optimized nominal current output, employing methods such as Bayesian optimization and a variety of acquisition functions. The effectiveness of the results is measured using accuracy, precision, recall, and F1 Score. By leveraging a comprehensive dataset spanning one year of METAR data, current consumption and utilizing Gaussian Process Regression for visibility predictions, in conjunction with K-Nearest Neighbor for determining the optimized nominal current output, this research reveals remarkable findings. The combination of these algorithms, implemented with holdout validation and expected-improvement acquisition function, achieves impressive results: 99.63% accuracy, 98.93% precision, 99.66% recall, and 99.29% F1 Score. The implications of this research are substantial, providing a clear pathway to optimizing nominal current output and, consequently, enhancing the overall efficiency and sustainability of airport operations.