Model-based trajectory generation starts from first principles. Given an aircraft type, a set of performance parameters, and a description of the environment, wind, airspace constraints, arrival slots, a trajectory can be computed by optimising some objective, such as minimising fuel consumption or ensuring separation from obstacles.
Well-established tools like OpenAP and BADA (Base of Aircraft Data) provide the aerodynamic and performance foundations for this kind of modelling. They can produce physically consistent, flyable trajectories, and they are already used across the aviation industry for simulation and procedure design.
The strength of this approach is also its main limitation: it is deterministic. Two runs with the same inputs produce the same trajectory. Real operations do not work that way. Pilots deviate from procedures, ATC issues vectors, wind behaves unexpectedly, and aircraft performance varies in ways that no parametric model fully captures. Attempts to introduce variability by adding random perturbations to the dynamics tend to produce distributions that do not match what is actually observed in radar data.
For VITOLMINS, model-based methods serve primarily as a baseline and a physical validation tool, a reference against which the outputs of data-driven approaches can be checked for plausibility. They are also valuable for generating trajectories in operational regimes where no historical data exists, such as novel VCA approach procedures that have never been flown.