Application to air sector clustering forecasting

After this introduction to various notions of artificial learning, we'll now look at how to apply learning methods to a concrete problem: forecasting air sector clusters. This problem is actually subdivided into two distinct sub-problems.

• The first is to model the air traffic controller's workload as a function of the traffic in the sector he controls. Each of these control sectors, also known as ATC (Air Traffic Control) sectors, is made up of one or more space sectors.

• The second sub-problem consists in finding an optimal partition of the space into ATC sectors, for each instant of the forecast horizon, so as to balance the workload and respect certain operational constraints, such as the number of available workstations.