A linear programming solution to the gate assignment problem at airport terminals

Abstract

This research solves the flight-to-gate assignment problem at airports in such a way as to minimize, or at least reduce, walking distances for passengers inside terminals. Two solution methods are suggested. The first is a heuristic algorithm which assigns the "most crowded" aircraft (i.e., most on-board passengers) to the best gate, while the second consists of formulating the problem as a linear program. A flight schedule of one day at Terminal No. 2 of Toronto International Airport is used to test and compare the two methods. The algorithm offers an assignment solution with a 27% reduction in the expected walking distance when compared to the original assignment at the airport. The linear program's assignment gives a 32% reduction. The heuristic algorithm is, therefore, only 5% suboptimal for the sample problem. In addition, its associated computational expenses, less than $10 per run, are by far cheaper than those of the linear program with expenses as high as $400 per run. Such excellent, or even acceptable, performance by the algorithm cannot be guaranteed for all problems. A strategy which helps decide when to use which approach is therefore suggested.