Top 10 most relevant articles in aviation
| Source | Objectives | AI models | Maintenance aspect | Short outcomes |
|---|---|---|---|---|
| Tseremoglou et al. (2023) | To optimize maintenance schedules by predicting the remaining useful life of components | DRL and MILP | Maintenance scheduling | The proposed MILP model produces more stable maintenance schedules and reduces maintenance ground time compared to other methods |
| Vos et al. (2023) | Maximize fleet availability while minimizing maintenance costs | RL | Maintenance scheduling | Demonstrated that RL-based policy can improve overall fleet availability, though the large action space in scheduling poses computational challenges |
| Ribeiro et al. (2022) | To develop interactive tools that enable human planners to design and visualize condition-based maintenance plans using ML support | ML integrated with a RL agent | Maintenance planning (condition-based maintenance) | Introduced a novel human–computer interaction approach that helps maintenance planners interact with an ML agent, leading to more reliable maintenance plans and better user understanding of AI-driven recommendations |
| Lee and Mitici (2023) | Leverage DRL and RUL prognostics for proactive aircraft maintenance decision-making | DRL combined with probabilistic RUL estimation | Predictive maintenance scheduling | Showed that a DRL approach can optimize maintenance timing by using RUL predictions, thereby reducing unscheduled downtime and improving maintenance efficiency |
| Tseremoglou and Santos (2024) | To schedule aircraft maintenance under uncertainty | DRL | Maintenance scheduling (condition-based maintenance) | Demonstrated that a deep RL algorithm can handle uncertain system information and still improve maintenance scheduling decisions, achieving higher availability under partial observability conditions |
| Dangut et al. (2022) | Improve fault prediction in aircraft systems by addressing class imbalance in maintenance data | BACHE algorithm with ML classifiers | Predictive maintenance | The proposed approach improved fault detection accuracy by balancing rare failure cases in the dataset, leading to more reliable identification of infrequent but critical failures |
| Kabashkin and Susanin (2024) | Support decisions on repair or replacement of aircraft components throughout their life cycle | Decision-support model (ML-based prognostics) | Maintenance planning | Provided a model that optimizes component maintenance and replacement timing over its life cycle, resulting in improved long-term fleet reliability and cost savings |
| Zhou et al. (2022) | Predict spare parts demands and plan component replacements by identifying competing failure risks in fleet equipment | Hybrid learning algorithm (statistical reliability models with ML) | Maintenance planning and spare parts management | Improves prediction of component failures and corresponding spare part needs, enabling efficient component replacement strategies and reduced downtime across the fleet |
| Rahim et al. (2020) | To minimize the life-cycle maintenance costs for critical aircraft components through optimized planning | AI-assisted optimization model | Maintenance planning | Developed a maintenance planning strategy that reduces overall costs for critical components by optimizing inspection and replacement intervals without compromising safety |
| Dang et al. (2024) | To enhance the transparency of AI-driven maintenance scheduling decisions using explainable AI techniques | Explainable AI integrated with maintenance scheduling algorithms | Maintenance scheduling (condition-based maintenance) | Demonstrated that providing causal explanations for AI-recommended maintenance schedules increases user trust and understanding, while still achieving efficient scheduling of aircraft maintenance |
| Source | Objectives | AI models | Maintenance aspect | Short outcomes |
|---|---|---|---|---|
| To optimize maintenance schedules by predicting the remaining useful life of components | DRL and MILP | Maintenance scheduling | The proposed MILP model produces more stable maintenance schedules and reduces maintenance ground time compared to other methods | |
| Maximize fleet availability while minimizing maintenance costs | RL | Maintenance scheduling | Demonstrated that RL-based policy can improve overall fleet availability, though the large action space in scheduling poses computational challenges | |
| To develop interactive tools that enable human planners to design and visualize condition-based maintenance plans using ML support | ML integrated with a RL agent | Maintenance planning (condition-based maintenance) | Introduced a novel human–computer interaction approach that helps maintenance planners interact with an ML agent, leading to more reliable maintenance plans and better user understanding of AI-driven recommendations | |
| Leverage DRL and RUL prognostics for proactive aircraft maintenance decision-making | DRL combined with probabilistic RUL estimation | Predictive maintenance scheduling | Showed that a DRL approach can optimize maintenance timing by using RUL predictions, thereby reducing unscheduled downtime and improving maintenance efficiency | |
| To schedule aircraft maintenance under uncertainty | DRL | Maintenance scheduling (condition-based maintenance) | Demonstrated that a deep RL algorithm can handle uncertain system information and still improve maintenance scheduling decisions, achieving higher availability under partial observability conditions | |
| Improve fault prediction in aircraft systems by addressing class imbalance in maintenance data | BACHE algorithm with ML classifiers | Predictive maintenance | The proposed approach improved fault detection accuracy by balancing rare failure cases in the dataset, leading to more reliable identification of infrequent but critical failures | |
| Support decisions on repair or replacement of aircraft components throughout their life cycle | Decision-support model (ML-based prognostics) | Maintenance planning | Provided a model that optimizes component maintenance and replacement timing over its life cycle, resulting in improved long-term fleet reliability and cost savings | |
| Predict spare parts demands and plan component replacements by identifying competing failure risks in fleet equipment | Hybrid learning algorithm (statistical reliability models with ML) | Maintenance planning and spare parts management | Improves prediction of component failures and corresponding spare part needs, enabling efficient component replacement strategies and reduced downtime across the fleet | |
| To minimize the life-cycle maintenance costs for critical aircraft components through optimized planning | AI-assisted optimization model | Maintenance planning | Developed a maintenance planning strategy that reduces overall costs for critical components by optimizing inspection and replacement intervals without compromising safety | |
| To enhance the transparency of AI-driven maintenance scheduling decisions using explainable AI techniques | Explainable AI integrated with maintenance scheduling algorithms | Maintenance scheduling (condition-based maintenance) | Demonstrated that providing causal explanations for AI-recommended maintenance schedules increases user trust and understanding, while still achieving efficient scheduling of aircraft maintenance |
Note(s): DRL: Deep reinforcement learning; MILP: mixed-integer linear programming; RL: reinforcement learning; ML: machine learning; RUL: remaining useful life; DSM: decision-support model; AI: artificial intelligence; CBM: condition-based maintenance; SVM: support vector machine
Sharing content requires targeting cookies to be enabled. Please update your cookie preferences to use this feature.