Optimizing predictive maintenance and mission assignment to enhance fleet readiness under uncertainty

Area of Science:

• Operations Research
• Reliability Engineering
• Asset Management

Background:

• Industrial fleets require maintenance during operational breaks.
• Existing Fleet Selective Maintenance (FSM) models lack realism by assuming identical missions and ignoring resource constraints.

Purpose of the Study:

• To develop a novel FSM model that jointly optimizes system-mission assignment, maintenance levels, and repair tasks.
• To integrate hybrid reliability assessment using analytical models and Deep Neural Networks (DNNs).
• To address uncertainties in maintenance and break durations via chance-constrained optimization.

Main Methods:

• A hybrid reliability assessment combining analytical models and DNNs for component health monitoring.
• A chance-constrained optimization model to ensure maintenance completion within break durations.
• Reformulation of the optimization model using Sample Average Approximation (SAA) and Conditional Value-at-Risk (CVaR) approximation.

Main Results:

• The proposed FSM model effectively integrates diverse component reliability data.
• The chance-constrained optimization ensures timely maintenance completion under duration uncertainties.
• The case study demonstrates the model's accuracy and practical value in military aircraft maintenance.

Conclusions:

• The novel FSM framework offers a more realistic and effective approach to fleet maintenance planning.
• Joint optimization of assignment, maintenance, and repair tasks significantly enhances fleet performance.
• The hybrid reliability assessment and uncertainty handling are crucial for real-world applications.