Intelligent simulation for alternatives comparison and application to air traffic management

Abstract

We present a simulation run allocation scheme for improving efficiency in simulation experiments for decision making under uncertainty. This scheme is called Optimal Computing Budget Allocation (OCBA). OCBA advances the state-of-the-art by intelligently allocating a computing budget to the candidate alternatives under evaluation. The basic idea is to spend less computational effort on simulating non-critical alternatives to save computation cost. In particular, OCBA is employed to intelligently provide the smallest number of simulation runs for a desired accuracy. In this paper, we present a new and more general OCBA scheme which can consider cases that users are interested not only the best design, but also any one in a good design set. In addition, this paper also presents the application of our OCBA to a design problem in US air traffic management. The national air traffic system in US is modeled as a large, complex, and stochastic network. The numerical examples show that the computation time can be reduced by 54% to 88% with the use of OCBA.