Model-Driven Maintenance Task-Personnel Matching Scheduling Method

The rapid development of the civil aviation industry has led to a continuous expansion in the scale of aircraft maintenance tasks that need to be decomposed. The design of rapid maintenance requirements for large-scale maintenance tasks can help improve maintenance efficiency. Existing maintenance task planning methods mostly focus on single-task optimization, ignoring the demand decomposition under large-scale tasks and the rapid matching of task requirements, making it difficult to meet the growing demand for aircraft maintenance tasks. This paper focuses on the matching and scheduling problem of large-scale maintenance tasks and personnel. Driven by deep learning models, a model-driven maintenance taskpersonnel capability scheduling method is proposed. This method first uses deep learning models to perform text classification on maintenance manual data, combining highfrequency actions with personnel capabilities. Then, the objective of minimizing the average time consumed per person is used for scheduling optimization to achieve automatic scheduling of tasks and personnel. Taking the disassembly task of Auxiliary Power Unit as an example, it is verified that this method improves scheduling efficiency while realizing automatic scheduling.