Crew cost decided at the pairing stage
Most crew expense is committed when pairings are built, not when rosters are published. Column-generation models search a space no manual process can.
The schedule, the crew, and the recovery.
Fleet assignment, crew pairing and rostering, and disruption recovery: the problems operations research was built on. We implement them against your network, your rules, and your union agreements.
Three problems worth modelling
Aircraft assigned to the demand actually there
Fleet assignment and tail routing determine both revenue capture and whether maintenance is legal downstream. They are one problem, not three.
Recovery in minutes, not in instinct
When a disruption cascades, the choice between swaps, cancellations, and reserves is a solvable problem. Modelled, it returns an answer inside the decision window.
What we build
- Fleet assignment matching aircraft type to demand across the network
- Tail routing that keeps maintenance opportunities legal and reachable
- Crew pairing via column generation under duty, rest, and agreement rules
- Crew rostering that respects bidding, seniority, and qualification constraints
- Disruption recovery for aircraft, crew, and passengers within the decision window
- Demand forecasting feeding schedule design and revenue management
Methods: column generation, MILP, network flows, constraint programming, heuristics
Outcomes
- Lower crew cost: pairings searched, not assembled
- Fewer illegal downstream states: maintenance reachability built into routing
- Faster recovery: a defensible plan while the disruption is still unfolding