Cluster & Disperse: a general air conflict resolution heuristic using unsupervised learning

TL;DR

A general and malleable heuristic for the air conflict resolution problem based on a new neighborhood structure for searching the solution space of trajectories and flight-levels and develops a novel algorithm for the horizontal plane based on a similar idea.

Abstract

We provide a general and malleable heuristic for the air conflict resolution problem. This heuristic is based on a new neighborhood structure for searching the solution space of trajectories and flight-levels. Using unsupervised learning, the core idea of our heuristic is to cluster the conflict points and disperse them in various flight levels. Our first algorithm is called Cluster & Disperse and in each iteration it assigns the most problematic flights in each cluster to another flight-level. In effect, we shuffle them between the flight-levels until we achieve a well-balanced configuration. The Cluster & Disperse algorithm then uses any horizontal plane conflict resolution algorithm as a subroutine to solve these well-balanced instances. Nevertheless, we develop a novel algorithm for the horizontal plane based on a similar idea. That is we cluster and disperse the conflict points spatially in the same flight level using the gradient descent and a social force. We use a novel maneuver making flights travel on an arc instead of a straight path which is based on the aviation routine of the Radius to Fix legs. Our algorithms can handle a high density of flights within a reasonable computation time. We put their performance in context with some notable algorithms from the literature. Being a general framework, a particular strength of the Cluster & Disperse is its malleability in allowing various constraints regarding the aircraft or the environment to be integrated with ease. This is in contrast to the models for instance based on mixed integer programming.

Figures (9)

• Figure 1: This figure shows a visual summary of the Cluster & Disperse heuristic.
• Figure 2: In the left figure, the conflict clusters are detected in the Clustering step. In the right figure, new tentative trajectories are shown after a planar CRP solver is used to resolve the conflicts.
• Figure 3: Two horizontal planes associated with two flight-levels are shown on the left. From each flight cluster, the top contributor to the conflict is removed and added to the other flight-level. The result is shown on the right. The newly formed trajectories on these two flight-levels have fewer conflicts.
• Figure 4: In this figure, four clusters related to one flight-level are shown. The conflict contribution score for each flight with respect to the cluster is shown next to them. As an example, we aim to choose $R=2$ flights in total while choosing $r=1$ top conflict-inducing flight from each cluster. These chosen flights are then assigned to another flight-level.
• Figure 5: RF-leg segments are arcs with fixed radius that connect two points. In this figure, a flight is instructed to fly over this arc to get from point O to D. The radius or the angle $\theta$ as will be used in our work gives an exact specification of the arc.