Accelerating stand demand modelling for hub airports: a Tabu search model as pre-solving approach

Abstract

Current airport stand demand modelling approaches show great potential in accurately determining stand demand, but face challenges with high computational times. In this paper, a novel pre-solving Tabu search model for determining the stand demand for hub airports is presented. The Tabu search model is designed to be used prior to a mixed-integer linear programming model such that it reduces computational time while retaining a close-to-optimal solution. It determines a stand demand based on a flight schedule. In this research, a case study is performed on two hub airports of comparable size. When comparing the results to the optimal stand demand, the combined model setup performs well in estimating the total number of stands, with median deviations of 3-5%; however, the estimations of the number of stands per stand type could be improved further. The Tabu search model ran for up to 40 minutes, and the combined model runtime was less than an hour in 78% of runs. It can be concluded that the stand demand can be estimated with reasonable accuracy using a combination of the Tabu search model and the exact model in less than an hour. However, the quality of the results remains limited in some cases; the primary recommendations are to test the model on multiple flight schedules across more airports and to improve the initial solution setup.