Pre-tactical time-slot assignment: Runway utilization and the impact of uncertainties

Abstract

Efficient planning of runway utilisation is one of the main challenges in Air Traffic Management (ATM). It is important because runway is the combining element between airside and groundside. Furthermore, it is a bottleneck in many cases. However, uncertainty and inaccuracy almost always lead to deviations from the actual plan or schedule. In this paper, we develop an optimization approach for the pre-tactical planning phase that provides some flexibility in the face of small disturbances in the input data, so that we need to change our plans less frequently. Instead of determining arrival/departure times to the minute in this phase yet, we assign several aircraft to the same time slot of a given size. The exact orders within those time windows can be decided later in tactical planning. Mathematically, this leads to a generalised assignment problem on a bipartite graph. We develop an integer program (IP), which can be solved very fast but may provide unnecessary large time buffers, and extend it to a mixed integer program (MIP) that solves the problem to global optimality. We present computational results concerning the abovementioned optimization approach and investigate the impact of disturbances on our deterministic solutions. As a next step, we will incorporate uncertainties directly in our model. Therefore, we analyse real-world data from a large German airport in order to obtain realistic delay distributions and describe a simulation environment to test current and future solution methods.