Local Redesigning of Airspace Sectors

In this paper we study the Airspace Sectorization Problem (ASP) where the goal is to find an optimal partition (sectorization) of the airspace into a certain number of sectors, each managed by an air traffic controller. The objective of the ASP is to find a “well-balanced” sectorization that distributes the workload evenly among the controllers. We formulate the ASP as a partitioning problem of a set of moving points in a polygonal domain. In addition to the requirement of balancing the workload, we introduce restrictions on the geometry of the sectorization which come from the Air Traffic Management aspects. We investigate several versions of the problem that arise from different definitions of the notion of the workload and various choices of geometric restrictions on the sectorization. We conclude that most of the formulations of the problem, except maybe in some trivial cases, are NP-hard. Finally, we propose a Local Redesigning Method (LRM), a heuristic algorithm that rebalances a given sectorization by adjusting the boundaries of the sectors. We evaluate LRM experimentally on synthetically generated scenarios as well as on the real historical traffic data. We demonstrate that the sectorizations produced by our method are superior in comparison with the current sectorizations of the US airspace.