Abstract
Flow line production systems consist of a number of stages (arranged in series) at which operations are performed on a work-piece. Operations at the stages are performed by machines or by equipment that are either perfectly reliable or subject to failure. While obtaining analytical results for the performance of a system with many machines, subject to failure, is considered to be an impossible task, also the approximate models are to be questioned. Approximate models are of two types: aggregation models or decomposition models. Assumptions have to be investigated on validity. By examining the up times and down times of the aggregated distribution by means of a discrete-event simulation model, the authors find that down times of aggregated machines in many situations do not follow an exponential distribution. An alternative distribution with the required characteristics is proposed and directions are given how to calculate its parameters.