Large-Scale Fleet Optimization Algorithms
In progressSince 2015
I designed a mixed-integer optimization architecture for scheduling a fleet of about 1,500 assets, using decomposition strategies and large-scale computational experiments to obtain exact solutions in under seven minutes. The work enabled near real-time re-optimization aimed at reducing downtime and improving fleet availability in asset-intensive operations.
Exact optimization engine for integrated flight scheduling and preventive maintenance planning at fleet scale.
Optimizes flight assignment and preventive maintenance timing simultaneously to maximize fleet availability while satisfying maintenance intervals, station capacity, and operational demand. Supports re-optimization when conditions change.
Decision Domain
Fleet flight scheduling + preventive maintenance planning (integrated operations & maintenance planning).
Methods
- Validated complex mathematical programming model
- Exact optimization algorithm
- Upper-bound algorithm design
- Computational experimentation for algorithm efficiency and effectiveness
Ideal Use Cases
- Airline / fleet flight + maintenance planning under station capacity constraints
- Multi-mission fleets with multiple maintenance types and intervals
- Organizations that need availability-maximizing schedules with traceable constraint handling
Inputs
- Fleet inventory (aircraft/assets and attributes)
- Mission types / flight demand requirements
- Maintenance activities with intervals and durations
- Maintenance station locations and capacity constraints
- Crew and operational constraints (as applicable)
- Residual flight time / usage-based counters (optional)
- Health indicators / prognostic signals (optional)
Outputs
- sub-7 minute solve time for a 1,500 fleet size
- Integrated flight + maintenance schedule plan
- Maintenance station loading plan and capacity utilization
- Projected availability over time (availability objective value)
- Constraint feasibility diagnostics and binding-constraint insights
- Re-optimization-ready model state (for updated conditions)