Ant Colony Optimization for Air Traffic Conflict Resolution

Abstract

The n aircraft conflict resolution problem is highly combinatorial and can be optimally solved using classical mathe- matical optimisation techniques only for small problems involving less than 5 aircraft. This article applies an Ant Colony Optimiza- tion (ACO) algorithm in order to solve large problems involving up to 30 aircraft. In order to limit the number of pheromone trails to update, a n aircraft conflict resolution problem is not modeled by a single ant but by a bunch of n ants choosing their trajectories independantly. A relaxation process is also used in order to be able to handle difficult conflicts for which partial solutions can help finding a path toward the optimal solution. Two different sizes of a toy problem are solved and presented.