A survey on open queueing network models applied to discrete manufacturing systems: Part I

In this paper the authors present the first part of the present survey on open queueing network models applied to discrete manufacturing systems. They review the so-called performance evaluation models to compute performance measures for open queueing networks, such as work-in-process, product leadtimes, production rate and equipment utilization. Exact and approximate decomposition methods are described for job-shops with multiple product classes and different workstations. To illustrate this, the authors present computational results from a network example of a semiconductor factory with 10 product classes and 13 workstations. They also discuss perspectives for future research. The second part of the present survey, to be presented in a subsequent paper, will review optimization models combining mathematical programming techniques and the theory of queueing networks. Roughly speaking, the difference between a performance evaluation model and an optimization model for queueing networks is that the first one describes performance measures under certain conditions while the second prescribes decisions.