So Many Factors, So Little Time... Simulation Experiments in the Frequency Domain

摘要: Abstract We illustrate how an efficient methodology called frequency domain experimentation can be used to gain better insight into the behavior of production systems. With full factorial designs commonly for simulation experiments, number runs grows exponentially with factors involved, while run length remains constant. In is independent factors, required lengths increase relatively slowly. describe method, its effectiveness at identifying important main effects, two-way interactions, and quadratic terms a known model, demonstrate approach by evaluating kanban system involving 34 provide links software. also present computational requirements running experiments that combine batch means run-oriented variety The results indicate perform very favorably systems, such as queueing networks, where simulation's output stream exhibits high positive autocorrelation.