Multi-agent based simulation for boarding

Abstract

Optimizing the boarding process is one way for the airline industry to reduce the cost of the total air plane turn time - i.e. the time between landing and take off. For this aim, boarding strategies that control when a passenger is allowed to board the plane are developed in order to reduce the overall boarding time. These strategies should be passenger friendly and robust to all kinds of disturbances. In this paper, boarding strategies are analyzed using multi-agent based simulation. Passengers are modelled as autonomous agents with particular individual characteristics. This allows us to introduce new strategies which are, unlike traditional boarding strategies, based on individual passengers' characteristics rather than on the location of their seats. We examine the robustness of these new and other common boarding strategies against realistic disturbances such as passengers compliance, group travel and transits. The effect of the latter two types of disturbance were not reported yet in literature. Our findings suggest that characteristic based strategies can combine high performance with low overhead.