Abstract: Compared with the traditional manufacturing system, several key characteristics make MRO(maintenance, repair and overhaul) scheduling problem different and difficult to handle, such as Disassembly-Repair-Assembly three-level structure, Material matching requirements, Stochastic routings and Variable processing times. On the basis of a detailed analysis of the MRO scheduling problem, a mixed integer linear programming model to minimize the expected total weighted tardiness is presented. Since this problem is NP-hard, a practical way of achieving good solutions is to develop effective simulation optimization algorithms. Multi-fidelity models, including one which is stochastic and time-consuming, are developed based on MO²TOS(multi-fidelity oftimization with ordinal transformation and optimal sampling) framework to solve this problem. Instead of directly running the high-fidelity model, a deterministic low-fidelity model is provided for OT(ordinal transformation) stage. Finally, the characteristics of different algorithms are analyzed by the numerical examples based on the practical background.

Key words: aviation maintenance, scheduling, stochastic routings, variable processing times, multi-fidelity models