OATAO - Open Archive Toulouse Archive Ouverte Open Access Week

An hybrid resolution method for the aircraft predictive maintenance and routing problem

Haloui, Ilyass and Ponzoni Carvalho Chanel, Caroline and Jimenez, Fabrice and Haït, Alain An hybrid resolution method for the aircraft predictive maintenance and routing problem. (2021) In: Scheduling and Planning Applications Wokrshop (SPARK) at ICAPS, 4 August 2021 (Virtual event, China). (Unpublished)

[img] (Document in English)

PDF (Author's version) - Depositor and staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader


Aircraft routing and maintenance planning still represents a challenging research topic. Aircraft routing, or aircraft tail assignment, optimises fleet allocation to predefined routes. Maintenance is optimized by accordingly making decisions based on aircraft preventive or condition-based needs. These two problems can lead to separate optimization processes. Nevertheless, a combined optimization problem should produce better results, as routes should impact maintenance costsand vice-versa. One issue is how to scale the problem such that many aircraft could fly at the same time from one airport to another and make maintenance decisions in an uncertain environment due to health indicators dynamics. In this work, we propose a unified model for this problem based on Markov Decision Processes (MDP). Two limiting factors impact our search of solution: the number of maintenance actions to be considered in planning that increases with the number of aircraft, as well as, the number of possible allocations that in turn is affected by the number of aircraft and airports. We propose a hybrid planning solution which decomposes the problem in such way it performs aircraft tail assignment optimisation combined with maintenance policy. The maintenance policies are computed for several independent MDPs (e.g. one per aircraft) in a robust manner by considering several possible aircraft locations. Then, a solution for the aircraft tail allocation problem is obtained by using aircraft maintenance policies. Finally, we empirically show that this method scales better than a unified (and possibly huge) MDP approach.

Item Type:Conference or Workshop Item (Paper)
Audience (conference):International conference without published proceedings
Uncontrolled Keywords:
Institution:Other partners > Airbus (FRANCE)
Université de Toulouse > Institut Supérieur de l'Aéronautique et de l'Espace - ISAE-SUPAERO (FRANCE)
Laboratory name:
Deposited On:25 Aug 2021 16:30

Repository Staff Only: item control page