The goal of the project is to develop efficient and accurate analytic/simulation methodologies to enable studies of auctioning airport runway slots and collaborative modeling of air traffic management. This new simulation paradigm is able to capture the dynamic and collaborative behavior of large, complex, stochastic networks.
Flight delays have steadily increased in recent years. In congested airports, this has lead to a compromise in safety, because the separation between aircraft has routinely dropped below the specified standard. In addition, many airlines are giving up larger jet airplanes and moving toward smaller regional jets, increasing the traffic at airports while reducing system capacity. Delays and canceled flights will only increase over time unless a new paradigm of air traffic management is adopted.
In this proposal, we consider using auctions to better regulate the use of runways for the best interest of the general public. To make effective policy decisions in this regard, accurate and efficient modeling of the air traffic network is necessary. The modeling must consider different viewpoints of the airlines, the airports, and the FAA Central Flow Control. Existing models are either purely analytical queueing models with many restrictive built-in assumptions or completely detailed simulation models with excessive computational requirements. To address this problem, there is a clear need for a technological breakthrough in methods of modeling air traffic management.
In this project we will develop new simulation methods using a hybrid simulation/analytic queueing model for capturing the dynamic behavior of large, complex networks, and efficiently run the simulation portion of the model using a new technique called Optimal Computing Budget Allocation being developed by the PI. Analytical queueing theory will be employed for aggregating nodes, with minimal loss in modeling accuracy, thus allowing a smaller network to be simulated. The budget allocation scheme is employed to provide the smallest number of simulation runs for a desired accuracy by not only considering variances of different alternatives as the standard theory now provides, but their relative profiles also. Establishing a robust and efficient simulation methodology will allow us to investigate runway slot auctioning and agent-based modeling (ABM) of air traffic management. ABM and auctioning can be viewed as operating on top of our fast simulation platform. Airlines and airports play the roles of agents engaged in a dynamic competition to maximize their individual profits with the use of information. Our fast simulation engine will enable us to experiment with alternative agent-based rules to determine optimal decision policies. Auctions have been successfully used for radio spectrum allocation with large numbers of interrelated regional licenses. However there are several challenging issues for designing runway slots auctions. For example, an airline's demand for a take-off slot at a flight’s originating airport is not independent of its demand for a landing slot at the flight’s destination airport. A poorly designed auction could lead to an unstable situation. Our fast simulation engine and ABM modeling approach will enable us to investigate and design good auction rules to better regulate the use of runways.
With the new development of the proposed simulation and modeling techniques, we will be able to efficiently analyze and provide effective solutions to a very complex system like the current U.S. air transportation network. Our modeling approach will allow us to experiment with alternative governmental constraints imposed on the air traffic system in order to maximize the public good. In particular, using auctions to allocate runway slots has the potential to economically induce airline carriers to more efficiently use the capacity of the entire system – thus, decreasing congestion and increasing the safety level.
In short, the major current research items include:
· Independent Verification & Validation Test on NASA VAMS Simulation Model
· Data Mining and Delay/Cancellation Space-Time Correlation Analysis
· Enroute Sector Workload Estimation
· Efficient Simulation and Optimization for Air Traffic Network
· Runway Slot Auctions for Demand Management
· Airport Safety Modeling & Analysis
This is a joint project with Professor George Donohue. The research is performed with Center for Air Transportation Systems Research at the Department of SEOR, GMU.
Back to Professor Chun-Hung Chen's Page