AM
A. Maniatis
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
1 records found
1
This thesis presents the development and evaluation of a scheduling framework for aircraft engine part repair tasks within the maintenance department of a major European airline. The study addresses inherent operational uncertainties, data limitations, and shortcomings of existing proactive scheduling practices. The proposed model extends the classic Resource Constrained Project Scheduling Problem (RCPSP) to one that combines multi-skill, maximum lags, and time-window constraints in a repair context. It employs a Tabu Search (TS) model with a custom lookahead decoder to increase feasibility, and integrates novel neighbourhood sampling techniques to handle large-scale operational data. The framework’s performance is assessed using a case study and sensitivity analysis to evaluate deadline adherence, schedule stability, and operational efficiency. Results demonstrate that an average tardiness objective offers the best trade-off between production efficiency and deadline adherence, and that its deadline-driven gradient yields schedules that stay robust as disruptions accumulate through the week. Different candidate lists for sampling based on the critical-path method perform on par with random sampling and show no statistically strong edge over it. Critical-path-based lists, however, rank top-two on every metric except stability and are never worst on any, supporting their use as a robust default. The findings highlight the interplay between a baseline and a repair scheduler and contribute to a more general understanding of scheduling in the repair context.
...
This thesis presents the development and evaluation of a scheduling framework for aircraft engine part repair tasks within the maintenance department of a major European airline. The study addresses inherent operational uncertainties, data limitations, and shortcomings of existing proactive scheduling practices. The proposed model extends the classic Resource Constrained Project Scheduling Problem (RCPSP) to one that combines multi-skill, maximum lags, and time-window constraints in a repair context. It employs a Tabu Search (TS) model with a custom lookahead decoder to increase feasibility, and integrates novel neighbourhood sampling techniques to handle large-scale operational data. The framework’s performance is assessed using a case study and sensitivity analysis to evaluate deadline adherence, schedule stability, and operational efficiency. Results demonstrate that an average tardiness objective offers the best trade-off between production efficiency and deadline adherence, and that its deadline-driven gradient yields schedules that stay robust as disruptions accumulate through the week. Different candidate lists for sampling based on the critical-path method perform on par with random sampling and show no statistically strong edge over it. Critical-path-based lists, however, rank top-two on every metric except stability and are never worst on any, supporting their use as a robust default. The findings highlight the interplay between a baseline and a repair scheduler and contribute to a more general understanding of scheduling in the repair context.