A simulation scenario based mixed integer programming approach to airline reserve crew scheduling under uncertainty
Bayliss, Christopher D. and de Maere, Geert and Atkin, Jason and Paelinck, Marc (2014) A simulation scenario based mixed integer programming approach to airline reserve crew scheduling under uncertainty. In: 10th International Conference on the Practice and Theory of Automated Timetabling (PATAT 2014), 26-29 August 2014, York, UK.
Airlines operate in an uncertain environment for many reasons, for example due to the efects of weather, traffic or crew unavailability (due to delay or sickness). This work focuses on airline reserve crew scheduling under crew absence and journey time uncertainty for an airline operating a single hub and spoke network. Reserve crew can be used to cover absent crew or delayed connecting crew. A fixed number of reserve crew are available for scheduling and each requires a daily standby duty start time. Given an airline's crew schedule and aircraft routings we propose a Mixed Integer Programming approach to scheduling the airline's reserve crew. A simulation of the airline's operations with stochastic journey time and crew absence inputs and without reserve crew is used to generate disruption scenarios for the MIPSSM formulation (Mixed Integer Programming Simulation Scenario Model). Each disruption scenario corresponds to a record of all of the disruptions in a simulation for which reserve crew use would have been beneficial. For each disruption in a disruption scenario there is a record of all reserve crew that could have been used to solve or reduce the disruption. This information forms the input to the MIPSSM formulation, which has the objective of finding the reserve schedule that minimises the overall level of disruption over a set of scenarios. Additionally, modifications of the MIPSSM are explored, and a heuristic solution approach and a reserve use policy derived from the MIPSSM are introduced. A heuristic based on the proposed Mixed Integer Programming Simulation Scenario Model or MIPSSM outperforms a range of alternative approaches. The heuristic solution approach suggests that including the right disruption scenarios is as important as ensuring that enough disruption scenarios are added to the MIPSSM.
Actions (Archive Staff Only)