Airport congestion and bottlenecks are a hassle for travelers and the airline industry. But new research from Maryland Smith is helping improve decision making within airport operations by producing accurate traveler forecasts in real-time.
Xiaojia Guo, assistant professor of decision, operations and information technologies at the University of Maryland’s Robert H. Smith School of Business, working alongside Yael Grushka-Cockayne and Bert De Reyck from the University of Virginia and Singapore Management University, is helping improve decision making within airport operations by producing accurate traveler forecasts in real-time.
The research, published in Manufacturing & Service Operations Management, is the first of its kind to apply machine learning with passenger-level data for journey forecasting. It features a two-phased predictive system that first anticipates the distribution of individual connection times and then forecasts the potential number of passengers arriving at immigration and security areas.
The first connection time predictions, they write, consist of the time difference between a passenger’s arrival at the airport and when they approach the airlines’ conformance desk. Their system then samples from the distribution of connection times to determine peak and trough passenger flow throughout the airport.
“Our predictive system 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,” write the researchers. “The airport can also update its resourcing plans based on the prediction of passenger flows.”
Guo and her fellow researchers partnered with London’s Heathrow Airport to address traveler connection times, which is the third most frequently filed passenger complaint.
In the past, once passengers had disembarked at their connections, airports possessed little knowledge of their whereabouts within the building – a contributing factor to delays in high-traffic areas. However, improving real-time passenger tracking capabilities enables airports to stabilize and predict departure times, as well as allocate resourcing needs more efficiently.
“Better predictions of the time passengers need to traverse the airport can help minimize such missed connections or, if unavoidable, can alert airlines in advance so that their schedules are not affected by unexpected late arrivals,” the researchers write. “Additionally, predictions of transfer passenger movements can help predict bottlenecks at immigration and security, allowing for preventive actions, thereby avoiding further delays.”
The method is currently being employed at Heathrow Airport and is outperforming the airport’s legacy systems in terms of identifying late passengers and predicting the number of passenger arrivals at security and immigration areas, according to the researchers. And it possesses great potential for application in other contexts, they write.
“Our predictive system can be generalized to other operations management domains, such as hospitals or theme parks, in which customer flows need to be accurately predicted,” write the researchers. “For future work, we encourage researchers and practitioners to formulate other decision-making problems that make use of distributional forecasts.”
Read More: “Forecasting Airport Transfer Passenger Flow Using Real-Time Data and Machine Learning,” in Manufacturing & Service Operations Management.
Media Relations Manager