A study of dynamic traffic re-routing in the National Airspace System

We exploit object technology to design a coarse grained model of an air traffic control system. Since aircraft delay is a critical variable in air traffic control, we consider two factors that contribute to delay: workload and capacity. Parameters to our model permit us to study workload by modifying takeoff frequencies at the airports under consideration. We study capacity by varying the frequency of arrivals and departures. Through modification of these parameters, the ripple effect of increased workload and reduced sector capacity on arrival and departure times is exposed. In both cases, we show that by dynamically rerouting traffic around problem sectors, we can reduce delays considerably; in some cases, dynamic rerouting reduced the number of delays by an order of magnitude. Incorporating the dynamic rerouting technique into our simulator required modifications to a single C++ class, showing that our object oriented model is easily modified and extended. However, a significant feature of our simulator is efficiency: we model a 24 hour period of air travel, including 50 airports with over 50,000 takeoffs, in a few minutes, obviating the need to parallelize our system.