Date of Award
Doctor of Philosophy
A goal of air traffic flow management is to alleviate projected demand-capacity imbalances at airports and in en route airspace through formulating and applying strategic Traffic Management Initiatives (TMIs). As a new tool in the Federal Aviation Administration's NextGen portfolio, the Collaborative Trajectory Options Programs (CTOP) combines many components from its predecessors and brings two important new features: first, it can manage multiple constrained regions in an integrated way with a single program; second, it allows flight operators to submit a set of desired reroute options (called a Trajectory Options Set or TOS), which provides great flexibility and efficiency.
One of the major research questions in TMI optimization is how to determine the planned acceptance rates for airports or congested airspace regions (Flow Constrained Areas or FCA) to minimize system-wide costs. There are two important input characteristics that need to be considered in developing optimization models to set acceptance rates in a CTOP: first, uncertain airspace capacities, which result from imperfect weather forecast; second, uncertain demand, which results from flights being geographically redistributed after their TOS options are processed. Although there are other demand disturbances to consider, such as popup flights, flight cancellations, and flight substitutions, their effect on demand estimates at FCAs will likely be far less than that of rerouting from TOSs. Hence, to cope with capacity and demand uncertainties, a decision-making under uncertainty problem needs to be solved.
In this dissertation, three families of stochastic programming models are proposed. The first family of models, which are called aggregate stochastic models and are formulated as multi-commodity flow models, can optimally plan ground and air delay for groups of flights given filed route choice of each flight. The second family of models, which are called disaggregate stochastic models and directly control each individual flight, can give the theoretical lower bounds for the very general reroute, ground-, and air-holding problem with multiple congested airspace regions and multiple route options. The third family of models, called disaggregate-aggregate models, can be solved more efficiently compared with the second class of models, and can directly control the queue size at each congested region. Since we assume route choice is given or route can be optimized along with flight delay in a centralized manner, these three families of models, although can provide informative benchmarks, are not compatible with current CTOP software implementation and have not addressed the demand uncertainty problem. The simulation-based optimization model, which can use stochastic programming models as part of its heuristic, addresses the demand uncertainty issue by simulating CTOP TOS allocation in the optimization process, and can give good suboptimal solution to the practical CTOP rate planning problem.
Airline side research problems in CTOP are also briefly discussed in this dissertation. In particular, this work quantifies the route misassignment cost due to the current imperfect Relative Trajectory Cost (RTC) design.
The main contribution of this dissertation is that it gives the first algorithm that optimizes the CTOP rate under demand and capacity uncertainty and is compatible with the Collaborative Decision Making (CDM) CTOP framework. This work is not only important in providing much-needed decision support capabilities for effective application of CTOP, but also valuable for the general multiple constrained airspace resources multiple reroutes optimization problem and the design of future air traffic flow management program.
Zhu, Guodong, "Decision making under uncertainties for air traffic flow management" (2019). Graduate Theses and Dissertations. 17631.