Robustness and evolutionary dynamic optimisation of airport security schedules
Reducing security lane operations whilst minimising passenger waiting times in unforseen circumstances is important for airports. Evolutionary methods can design optimised schedules but these tend to over-fit passenger arrival forecasts resulting in lengthy waiting times for unforeseen events. Dynamic re-optimisation can mitigate for this issue but security lane schedules are an example of a constrained problem due to the human element preventing major modifications. This paper postulates that for dynamic re-optimisation to be more effective in constrained circumstances consideration of schedule robustness is required. To reduce over-fitting a simple methodology for evolving more robust schedules is investigated. Random delays are introduced into forecasts of passenger arrivals to better reflect actuality and a range of these randomly perturbed forecasts are used to evaluate schedules. These steps reduced passenger waiting times for actual events for both static and dynamic policies with minimal increases in security operations.
Citation : Chitty, D.M., Yang, S. and Gongora, M. (2017) Robustness and evolutionary dynamic optimisation of airport security schedules. Proceedings of 23rd International Conference on Soft Computing (MENDEL 2017), pp. xxx-xxx
Research Group : Centre for Computational Intelligence
Research Institute : Institute of Artificial Intelligence (IAI)
Peer Reviewed : Yes