Computing budget allocation for simulation experiments with different system structures

Abstract

Simulation plays a vital role in analyzing discrete event systems, particularly in comparing alternative system designs with a view to optimize system performance. Using simulation to analyze complex systems, however, can be both prohibitively expensive and time consuming. We present effective algorithms to intelligently allocate computing budget for discrete event simulation experiments with different system structures. These algorithms dynamically determine the best simulation lengths for all simulation experiments and thus significantly reduce the total computation cost for a desired confidence level. Numerical illustrations are included. We also compare our algorithms with our earlier approach in which different system structures are not considered. Numerical testing shows that we can further improve simulation efficiency.