Xiaojia Guo - Heathrow

Forecasting Airport Transfer Passenger Flow Using Real-Time Data and Machine Learning

Xiaojia Guo (Robert H. Smith School of Business, UMD), Yael Grushka-Cockayne (Harvard Business School), Bert De Reyck (Lee Kong Chian School of Business, Singapore Management University)

Airports have been challenged to improve collaborative decision-making by producing accurate forecasts in real time. In collaboration with Heathrow, we develop a two-phased predictive system that produces forecasts of transfer passenger flows. In the first phase, the system predicts the entire distribution of transfer passengers’ connection times. In the second phase, the system samples from the distribution of individual connection times and produces distributional forecasts for the number of passengers arriving at the immigration and security areas. The predictive system developed is based on a regression tree combined with copula-based simulations. We generalize the tree method to predict complete distributions, moving beyond point forecasts. Theoretically, we show that point forecasts of passenger flows generated by the two-phased approach are unbiased, and that distributional forecasts can be well-calibrated when adding correlation as a tuning parameter. When compared to benchmarks, our two-phased approach is shown to be more accurate in predicting both connection times and passenger flows. Our predictive system has been implemented at Heathrow since 2017. It can produce accurate forecasts, frequently, and in real-time. With these forecasts, an airport’s operating team can make data-driven decisions, identify late passengers and assist them to make their connections. The airport can also update its resourcing plans based on the prediction of passenger flows. In a back-testing simulation study, we show that compared to several benchmark methods, the resourcing decisions made by applying predictions produced by our predictive model could better utilize resources and reduce passenger delays. Although our approach is developed for managing airport passenger flows, it can be generalized to other operations management domains, such as rail or theme parks, in which both arrival times and number of arrivals need to be accurately predicted.

(Manufacturing & Service Operations Management 2022; Winner of 2022 INFORMS Aviation Applications Best Paper Award; Finalist of 2018 MSOM Practice-based Research Competition)

(a) Pasenger Arrivals Generated by a Time Series Model (TBATS)

(b) Pasenger Arrivals Generated by the Two-phased Model

Figure 1. Number of Passenger Arrivals at Immigration and Security Areas on Test Day December 4, 2015. It is easy to observe the accuracy of the forecasts from the two-phased model.

Figure 2. Output files of the Application Developed for Heathrow to Use the Two-phased Approach in Real Time. In addition to individual paseengers' probability of being late and passenger flows at immigration and security areas, we also provide the expected number of late passengers for each outbound flight.