Aerospace and Electronic Systems - May 2019 - 13

Image credit: The Authors

optimization methods have yet to be applied to the whole
airport. The contribution of this paper is therefore the
application of our optimization algorithm to the whole airport decision problem. An additional contribution of our
research is a comparison of the results of our algorithm
with the performance of human air traffic controllers
(ATCOs) using the same data; such an analysis has not yet
been performed in the context of the cited literature.
The objective of this paper is to investigate the usefulness of a holistic optimization approach for ATM at an
airport using a realistic case study. Holistic here has
multiple meanings. First, we address the entire traffic
control problem, which includes routing, sequencing, and
scheduling airplanes on the surface, and sequencing and
scheduling airplanes on the runway. Second, our approach
goes beyond the natural decomposition process performed
by ATCOs, by considering the entire problem in space
and time. Indeed, each controller takes care of only a
small part of the airport (i.e., his/her area of responsibility)
and controls a few airplanes at a time (i.e., the next ones in
his/her area/sector). In our case study, we focus on the second point by comparing the performance of our methodology with that of human operators.
To this end, we developed and evaluated a software
tool for optimizing the decision-making process at an
airport in a holistic way and applied it to the whole airport
decision problem at Hamburg airport. In previous work,
[9]-[11] we proposed a productivity improvement
process, called Zero Failure Management at Maximum
Productivity in Safety Critical Control Rooms (ZeFMaP),
to increase the performance within control rooms. Our
results showed that ZeFMaP could indeed help improve
productivity in the context of tower control rooms. In
[12], we describe in full detail the mathematical
background for the optimization algorithm applied in our
case study, and briefly report the major computational
MAY 2019

results. In this paper, we focus on the impact of such holistic optimization on the overall decision process and performance of air traffic control at the airport. To better
understand how the algorithm can support ATCOs in their
work, we analyzed the ATCOs self-reported descriptions
of the current decision-making process.

HOLISTIC OPTIMIZATION ALGORITHM
Our holistic optimization algorithm does not look at each
airplane trajectory separately but rather at all possible
airplane trajectories together, to determine the globally
optimal decision at each point in time for each airplane.
Note that trajectories are not only defined through airplane
(i.e., spatial) movements on the surface at the airport and on
the runway but also through the timing associated with these
movements (i.e., temporal). A solution for all considered
airplanes is called a plan.
We are defining airplane movements based on the
airport graph, which provides a schematic representation
of the airport. Constraints decide whether a plan or a
combination of trajectories is feasible. In addition to single business trajectory constraints (turning restrictions on
the taxiway and speed limitations), there are constraints
relating to pairs of trajectories (separation constraints).
Our model obeys the same safety rules (runway separations, airplane turning restrictions, etc.) as those that
ATCOs are instructed to follow during training. The
objective used by the algorithm is to choose the best
possible plan. Possible defined goals include minimizing
total taxi time and maximizing punctuality.
Note that in our exercise, ATCOs received in advance a
list of arrival times and desired departure times. In the simulations, airplanes landed precisely at the communicated arrival
times, and no deviations occurred. Accordingly, our model

IEEE A&E SYSTEMS MAGAZINE

Aerospace and Electronic Systems - May 2019

Table of Contents for the Digital Edition of Aerospace and Electronic Systems - May 2019