Reduction of air traffic congestion by genetic algorithms

Sofiane Oussedik (1) , Daniel Delahaye (2)

1 CMAP - Centre de Mathématiques Appliquées
2 ENAC - Ecole Nationale de l'Aviation Civile

The annual number of flights in Western Europe has increased from about 2.6 million in 1982 to about 4.5 million in 1992, an increase of 73%. Acute congestion of the Air Traffic Control system has been the result. One way to reduce this congestion is to modify the flight plans (slot of departure and route) in order to adapt the demand to the available capacity. This paper addresses the general time-route assignment problem. A state of the art of the existing methods shows that this problem is usually partially treated and the whole problem remains unsolved due to the complexity induced. We perform our research on the application of stochastic methods on real traffic data, and without using the flow network concept, but by simulating the flight of each aircraft. The first results shows that our Genetic Algorithms based method is able to reduce congestion of the french airspace by a factor 2. Special coding techniques and operators are used to improve the quality of the genetic search.

Domaines: Optimisation et contrôle [math.OC]

Sofiane Oussedik, Daniel Delahaye. Reduction of air traffic congestion by genetic algorithms. Lecture Notes in Computer Science, 1998, Lecture Notes in Computer Science, 1498, pp 855-864. ⟨10.1007/BFb0056927⟩. ⟨hal-01018427⟩