Predicting block times: An application of Box-Jenkins models

Predictability is an important key performance area because it influences airlines’ on-time performance and passengers’ satisfaction. Predictability also plays a significant role in ensuring a successful transition from air traffic control to a managed air traffic system. Yet, predictable block times are all the more difficult to achieve as the on-time execution of a flight may be constrained by three types of delay which airlines cannot control: induced, propagated and random. This paper proposes a methodology based on the Box-Jenkins models to measure the predictability of block times as an alternative to variance or standard deviation of block delays. The padding of schedules and the timeliness of flight plans are likely to bias the variation in block delays as an indicator of predictability.