Machine learning for predicting the impact of construction activities on air traffic operations during airport expansion projects

Abstract

Construction activities during airport expansion projects disrupt air traffic operations and often need to be performed in phases to minimize their disruptive impacts. This paper presents a machine learning methodology for quantifying the impact of alternative construction phasing plans on air traffic operations. The methodology is implemented in four stages: data collection, data preprocessing, model training, and evaluation stages. A case study is analyzed to highlight the original contributions of the methodology that include (1) development of five machine learning models for accurately and efficiently quantifying the impact of construction-related airport closures on flights ground movement time, (2) comparison of the performance and prediction accuracy of the developed models, and (3) efficient assessment of the impact of alternative construction phasing plans on airport operations without the need for time-consuming simulations. This is expected to provide planners with much-needed support to identify construction phasing plans that minimize total flight delays.