Swarm Theory Applied To Air Traffic Flow Management

Abstract

The goal of air Traffic Flow Management (TFM) is to balance demand against capacity in order to reduce inefficiencies. As an optimization problem TFM poses a number of difficult challenges. From the airline perspective the solutions should minimize the aggregate delay time relative to the scheduled time that is used by airlines to drive their operations. TFM must take into account air traffic control restrictions used to maintain aircraft separation and changes in capacity due to weather disruptions. In current operations TFM is done based on a centralized approach that relies on predictions and that does not integrate airline preferences. Combinatorial optimization techniques to solve the multi-objective traffic flow optimization problem are not practical; the vast number of variables and the exceedingly large Pareto front associated with the solution space generates a combinatorial explosion that makes the approach completely intractable. This paper presents a different approach to TFM, inspired in swarm theory, that converts pilots into goal seeking agents that individually find local solutions to the optimization problem and as a whole the collective action of agents creates emergent behavior that naturally tends to converge on its own to a Pareto efficient state.