Optimization of Ramp Area Operations

Airport runways and taxiways have been identified as a bottleneck of the national airspace system, and the major inhibiting factor for serving an increasing air traffic demand. In order to keep up with the increase of aircraft density, new techniques are required to increase airport throughput while maintaining safe separation constraints. Since key airports that accommodate a large portion of traffic operate at or close to their maximum capacity, an optimization of runway and taxiway operations is necessary.

Most of the prior research on taxiway scheduling has focused on modeling an airport as a graph, i.e., a connected network, with aircraft traveling along the graph edges. In order to solve the optimization problem on the graph authors have used genetic algorithms, Mixed Integer Linear Programs (MILPs), hybrids of these, constrained search algorithms, and generalized dynamic programming algorithms. Most of the previous work, however, does not account for the planning and execution of ramp area operations. Although the surface operations can be improved by adopting an optimal taxiway schedule, its execution ultimately depends on human controllers who control aircraft maneuvers in both ramp area and taxiways. Ramp area aircraft have been incorporated in some prior works, but the trajectories are considered to be deterministic.

The main difficulty in the integration of ramp area operations into an optimal taxiway scheduling solution is in addressing uncertainties of ramp area trajectories. Unlike aircraft maneuvers on taxiways, ramp area aircraft maneuvers are typically not confined to well-defined trajectories. The shape and timing of the trajectories are subject to uncertainties resulting from pilots decisions as well as other factors involved in ramp area operations, which can impede upon an optimal taxiway schedule plan. Ramp area aircraft have been incorporated in prior work, but the trajectories are considered to be deterministic.

To address the uncertainty of ramp area operations, we proposed a stochastic model of ramp area aircraft trajectories. The model was used to sample a large number of feasible ramp area aircraft trajectories at the Dallas Fort Worth Terminal C ramp area. Using the sampled trajectories, we computed conflict distributions among any two aircraft defined by their relative taxiway spot schedules on the graph. The conflict distributions were used to estimate conservative conflict separation constraints that were passed to a MILP built from the Spot and Runway Departure Advisor (SARDA) logic. A similar method was also applied to the center alley of the Charlotte Douglass ramp area

During time periods of heavy traffic, SARDA advises departing aircraft to remain at the gate with engines off until just before their scheduled spot release time, and when cleared, they can proceed straight from the taxiway spot to the departure runway queue without slowing down or stopping for other traffic. This technique has the effect of significantly reducing fuel burn and engine emissions. Recent studies have estimated as much as 18% of fuel consumption during taxi operations was due to stop-and-go activity. For departing aircraft to proceed along the route unimpeded, it is critical that the aircraft arrive at the taxiway spot at the scheduled time. It is assumed that airports have the necessary tools to meet the assigned taxiway spot release times.

In addition to computing the separation constraints, this research provides a tool to aid ramp area controllers in meeting the scheduled times by computing the feasible push back time window for each departing aircraft. The push back window is defined by the range between the earliest feasible push back time and latest feasible push back time. Initiating the push back within these bounds ensures there exists a feasible trajectory that arrives at the taxiway spot at the required time, as required by the optimal schedule. For relative taxiway schedules that are conflict free, computing the push back windows is straightforward and can be estimated from the sampled trajectories. For relative schedules with a non-zero ratio of conflicts, an optimization procedure is need to solve for the optimal combination of push back time windows that are constrained to be conflict free in the presence of trajectory uncertainties.