This dissertation seeks to design an optimization algorithm, based on naturalistic flight data, with emphasis on safety to perform a benefits' analysis when sequencing and scheduling aircraft at the runway. The viability of creating a decision-support tool to aid air traffic controllers in sequencing and optimizing airport operations is evaluated through the benefits' analysis. Air traffic control is a complex and critical system that ensures the safe and efficient movement of aircraft within the airspace. This is particularly true in the immediate vicinity of an airport. Unlike in en-route or terminal area airspace where aircraft usually traverse well established routes and procedures, near the airport after completing a standard arrival procedure, the routes to the final approach are only partially defined. With safety being the foremost priority, the local tower controllers monitor and maintain separation between aircraft to prevent collisions and ensure the overall safety of the airspace. This involves constant surveillance, coordination, and decision-making to manage the dynamic movement of aircraft, changing weather conditions, and potential hazards. All the while, the controllers make decisions regarding tromboning or vectoring based on various factors, including traffic volume, airspace restrictions, weather conditions, operational efficiency, and safety considerations to ensure a safe traffic sequencing of aircraft at the runway. A novel framework is presented for modeling, characterizing, and clustering aircraft trajectories by extracting traffic control decisions of air traffic controllers. A hidden Markov model was developed and applied to transform trajectories from a sequence of temporal spatial position reports to a series of control actions. The edit distance is utilized for quantifying the dissimilarity of two variable-length trajectory strings, followed by the application of k-medoids algorithm to cluster the arrival flows. Next, a repeatable process for detecting and labeling outlier trajectories within a cluster is introduced. Through application on a set of historical trajectories at Ronald Reagan Washington National Airport (DCA), it is demonstrated that the proposed clustering framework overcomes the deficiency of the classical approach and successfully captures the arrival flows of trajectories, that undergo similar control actions. Leveraging on the set of arrival flows, statistical and machine learning models of air traffic controllers are created and evaluated when ordering aircraft to land at the runway. The potential inefficiencies are identified at DCA when sequencing aircraft. As such, there is a potential performance gap, and it appears that there is room for additional sequence optimization. With the goal of overcoming the potential inefficiencies at DCA, a mixed-integer zero-one formulation is designed for a single runway that takes into consideration safety constraints by means of separation constraints between aircraft imposed at each metering point from the entry to the airspace until landing. With the objective of maximizing runway throughput and minimizing the traversed distance, the model sequences and schedules arrivals and departures and generates safe and conflict-free arrival trajectories to actualize that scheduling. The output of the optimization shows that the model successfully recovers approximately 52% of the performance gap between the actual distance traversed and idealized (cluster centroids) distance traversed by all arrival aircraft. Moreover, each arrival aircraft, on average, traverses 2.12 nautical miles shorter than its historical trajectory and thus saving approximately 10 US gallons of jet fuel. By showcasing the potential benefits of the optimization, this dissertation takes a step towards achieving the long-term vision of developing a decision-support tool to assist air traffic controllers in optimally sequencing and scheduling aircraft. To fully leverage the potential benefits of optimization, further development and refinement of the algorithm are necessary to align it with real-world operational demands. As future work, the research would be expanded to integrate uncertainties like weather conditions, wind directions, etc. into the optimization.


If this is your thesis or dissertation, and want to learn how to access it or for more information about readership statistics, contact us at STARS@ucf.edu

Graduation Date





Vela, Adan


Doctor of Philosophy (Ph.D.)


College of Engineering and Computer Science


Industrial Engineering and Management Systems

Degree Program

Industrial Engineering


CFE0009703; DP0027810





Release Date

August 2023

Length of Campus-only Access


Access Status

Doctoral Dissertation (Open Access)