Pruning Rules for Optimal Runway Sequencing

This paper investigates runway sequencing for real-world scenarios at one of the world’s busiest airports, London Heathrow. Several pruning principles are introduced that enable significant reductions of the problem’s average complexity, without compromising the optimality of the resulting sequences, nor compromising the modeling of important real-world constraints and objectives. The pruning principles are generic and can be applied in a variety of heuristic, metaheuristic, or exact algorithms. They could also be applied to different runway configurations, as well as to other variants of the machine scheduling problem with sequence dependent setup times, the generic variant of the runway sequencing problem in this paper. They have been integrated into a dynamic program for runway sequencing, which has been shown to be able to generate optimal sequences for large-scale problems at a low computational cost, while considering complex nonlinear and nonconvex objective functions that offer significant flexibility to model real-world preferences and real-world constraints. The results shown here clearly demonstrate that, by exploiting the problem structure, complex runway sequencing problems can be solved exactly.