A methodology for the design of simulation campaigns based on population variance characterization

Manufacturing industrial systems are complex systems whose performance is characterized by interactions among different parts of the system as well as by stochastic phenomena affecting the operation of the parts themselves.A key aspect in studying a complex system is the ability to model its evolution over time and, as a consequence, to identify, from a statistical point of view, the trend of the performance measures (i.e. productivity) over time. Discrete event system simulation (DESS) is certainly the widespread technique adopted to this aim.In this paper, a methodology to characterize the trend of the variance of the population for a flow-line production system is developed. The knowledge of the relation between the variance of the population and the system run time allows the analyst to better design simulation campaigns and define warm-up period. Moreover, this result is also useful when in-field tests have to be designed to certify performances of a newly deployed system