FLEET is equipped with powerful optimisers capable of obtaining optimised solutions for large scheduling and rescheduling problems

• Routing scheduling
Schedules more than 10 000 train trips in problems where all train compositions are homogeneous (they don’t mix different vehicle types) and more than 1 500 train trips in problems where train compositions are heterogeneous and have more than 20 different possible configurations
• Roster scheduling
Schedules more than 1 000 train trips in more than 300 lines of multiple cyclic rosters of 15 different vehicle types
• Short-term scheduling
Reschedules more than 150 individual calendar schedules of 15 different vehicle types for a period of 20 days, which means allocating or reallocating more than 1 000 train trips to those calendar schedules
• Real-time dispatching
Allocates over 200 individual calendar schedules of 15 different vehicle types for a period of 2 days


FLEET optimisers obtain solutions that optimise several goals at the same time. Popular, but sometimes conflicting goals are:

• Routing scheduling
Minimise operational costs (e.g. energy consumption, maintenance costs, track slot allocation, crew approximate costs, vehicle purchasing costs, and mismatches between preferred and allocated vehicle types); maximise passenger satisfaction (e.g. seat availability), revenue (e.g. seats sold taking into consideration passenger demand and train capacity) and robustness (e.g. by avoiding shunting operations)
• Roster scheduling
Similar as in routing scheduling, but with much more accurate evaluation of maintenance and purchasing costs
• Calendar scheduling
Minimise the number of changes with respect to the original solution and the number of trains with no allocated equipment, in addition to the previous goals
• Real-time dispatching
Minimise recovery time, in addition to the previous goals


FLEET optimisers use state-of-the-art technology, namely column generation, Lagrangian relaxation, integer linear programming, network flow algorithms, dynamic programming and several metaheuristics.
They are integrated in a very user-friendly way that allows the user to run several what-if scenarios just by changing parameter values. FLEET optimisers take advantage of the structural properties of the problem. For instance, problems where the output is routing schedules (with no maintenance constraints), and where train compositions are homogeneous, can be solved very quickly with network flow algorithms.