Synchronizing transportation of people with reduced mobility through airport terminals

Abstract

Navigating through an airport is easy enough for most passengers, but when you are reduced in mobility it is a different story. In this paper we look at an airport at which between 300 and 500 Passengers with Reduced Mobility (PRMs) arrive daily, who need assistance from airport employees in their journey. We want to find a schedule for these employees to support as many PRMs as possible while ensuring a smooth journey with little waiting time. PRMs may not be left unsupervised, except in a lounge. Since the employees are only allowed to work within their own terminal, the tasks of the employees must be synchronized to enable a smooth hand-over of a PRM. Moreover, we want to find a robust schedule to be resistant against minor disturbances. Since we must be able to reschedule in case of a major disturbance, the maximum computation time of our algorithm is restricted to two minutes. This problem was first studied by Reinhardt et al., who present an insertion heuristic, where the order in which the PRMs are considered is modified through local search. We present a decomposition model in which we first determine feasible start times for the tasks describing the journeys of the passengers using Simulated Annealing, after which in each iteration we assign the tasks to the employees using a (heuristic) matching algorithm. Experimental results show that our algorithm is able to ensure smooth, robust connections, while supporting every passenger in the given deterministic instances, which originate from Reinhardt et al. Finally, we present a simulation study to test our approach in a dynamic environment. It turns out that it can easily deal with these disturbances in real time and come up with very good solutions. The code used in the simulation study is freely available at https://github.com/RvanTwist/uu-thesis-prm.