State-Aware Phased System Update Strategy for Cooperative Scheduling of Fleet Missions

Frequent OTA updates in large commercial fleets must be planned around limited ground windows and strict safety requirements. Our study develops a phased OTA strategy that adapts update decisions to aircraft operational states. Real-time ACARS/ADS-B data and maintenance logs were combined to train an LSTM-Attention model for predicting remaining ground time, achieving an RMSE of 4.7 minutes. An integer-programming scheme then assigned update granularity and rollback plans across short stops, overnight stays, and scheduled A/B/C checks. Experiments using six months of data from 86 narrow-body aircraft showed that critical-system updates kept the impact on on-time performance within 0.3%, increased ground-time utilization by 21.8%, and reduced cross-version maintenance reports by 44.5%. The results indicate that state-aware OTA scheduling can support safer and more efficient update cycles in operational fleets.