Modeling and control of airport queueing dynamics under severe flow restrictions

Abstract

Based on field observations and interviews with controllers at BOS and EWR, we identify the closure of local departure fixes as the most severe class of airport departure restrictions. A set of simple queueing dynamics and "traffic rules" are developed to model departure traffic under such restrictions. The validity of the proposed model is tested via Monte Carlo simulation against 10 hours of actual operations data collected during a case-study at EWR on June 29,2000. In general, the model successfully reproduces the aggregate departure congestion. An analysis of the average error over 40 simulation runs indicates that flow-rate restrictions also significantly impact departure traffic; work is underway to capture these effects. Several applications and "what-if'' scenarios are discussed for future evaluation using the calibrated model. grams to congested airports). Note that both types of restrictions have an effective cutoff threshold, above which traffic is stopped completely. Restrictions can also be categorized by spatial scope: This paper focuses on successfully modeling the most significant ("first-order") effects of downstream restrictions, where the relative severity is measured in terms of overall impact on aggregate metrics such as airport throughput, departure congestion, and average taxi-out delay.

Network Congestion Control of Airport Surface Operations

“…[4][5][6][7] Since surface operations are highly heterogenous in terms of taxi speeds, routes and pilot behavior, Eulerian models 8,9 of traffic flow are not applicable here. Most traditional solution approaches involve the use of optimal scheduling algorithms 6,7 or queuing theory, [3][4][5] both of which have certain drawbacks. Optimization using linear or mixed-integer programming assumes that aircraft move at constant velocities at all times and follow time-based taxi instructions exactly, which are not realistic assumptions given the current state of technology at airports.…”

A network congestion control approach to airport departure management

“…The projected traffic level, k p , can be calculated based on the expected times between successive departures, as given in Eqn. (6). Since entry into the network is assumed FCFS, the projected traffic level decreases as t p increases (there can be no additional aircraft entering the network while the current aircraft is waiting).…”

“…Most traditional approaches to this problem involve the use of optimal scheduling algorithms [2], [3] or queuing theory [4], [5], [6]. Optimization methods typically assume that aircraft move at constant velocities at all times, and follow time-based taxi instructions exactly.…”

“…As this pushback rate approaches the maximum throughput, large queues are formed at the departure runway, in addition to conflicts happening at other points on the surface. While there have been several studies that aim to relieve surface congestion [1], [2], [3], [4], [5], [6], they typically either fail to consider the inherent stochasticity of aircraft movement and pilot behavior, or are intrinsically numerical in nature. This work presents a formulation that explicitly accounts for stochastic taxi-out behavior, while providing insight into the system via an analytical treatment of airport performance.…”

Software tools to support research on airport departure planning

“…For applications requiring finer detail than ASQP provides, good results have been obtained using data from FAA tower flight-strips [2]. When combined with tower and TRACON logs, flightstrip data provides a very thorough and complete description of traffic at a given airport.…”