Applying statistical control techniques to air traffic simulations

Abstract

While the literature contains several adaptive sampling techniques for statistical comparison of competing simulated system configurations and for embedded statistical computations during simulation run-time, these techniques are often difficult to apply to air traffic simulations because of the complexity of air traffic scenarios and because of the variety of model and data types needed to fully describe air traffic. Adaptive sampling techniques can be beneficial to the study of air traffic; for example, adaptive techniques can use ranking and selection methods to compare the relative worth of the competing configurations and calculate the number of observations required for rigorous statistical comparison, often dramatically reducing the run-time duration of simulations. In this paper, we will describe the implementation of such procedures in the Reconfigurable Flight Simulator for air traffic simulations. We also discuss implications for the coordination of simulation, analysis, and design activities.