EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

4052. Optimizing Aircraft Line Maintenance Scheduling: A Reinforcement Learning Approach

Invited abstract in session WD-26: Combinatorial optimization issues in transportation (Contributed), stream Combinatorial Optimization.

Wednesday, 14:30-16:00
Room: 012 (building: 208)

Authors (first author is the speaker)

1. Syed Shaukat
Aviation, University of New South Wales

Abstract

In the dynamic and complex realm of the airline industry, efficient and reliable scheduling and execution of aircraft maintenance are crucial. Significant challenges in the optimal scheduling of aircraft line maintenance arise from the need to balance maintenance requirements with the limited availability of resources and dynamic operational constraints. Traditional approaches typically employ optimization-based methods to address these issues. However, these methods often struggle to scale up in real-world applications, along with the uncertainties and non-linearities prevalent in such scenarios. Recent advancements in airline scheduling signal a paradigm shift, with simulation-based, data-driven, and AI-enhanced approaches superseding conventional optimization models. This paper introduces an innovative approach that capitalizes on machine learning by integrating a hybrid simulation model with a reinforcement learning (RL) algorithm and dynamic programming search.

Utilizing a hyper-heuristic learning approach, the RL agent explores and learns the optimal scheduling of maintenance tasks, encompassing a criterion-based action space and considering various factors such as aircraft availability, resource allocation, and maintenance requirements. The results surpass the performance of existing optimization-based methods and offer valuable insights that may be applicable to other industries facing similar scheduling challenges in the transportation and supply chain sectors.

Keywords

Status: accepted

Back to the list of papers