Stochastic analysis of versatile workcentres

We analyse the system consisting of a highly capable workcentre, which processes a variety of part types, using queueing models. The various part types produced by the system have distinct arrival and processing durations that are stochastic in nature. When an arriving workpiece finds the machine busy, it waits in a pre-process storage buffer (queue); this buffer may be common for all the part types, or may be dedicated for that part type. Further, this buffer may be capable of holding only a finite number of workpieces, or may be of infinite capacity. When the machine changes over from producing one type of part to another, a setup operation of stochastic duration is necessary to adjust the machine and load the necessary tools for production of the next part type. This model is representative of a typical machining centre in an Automated Manufacturing System. We focus on GI/G/1 models and multiqueue polling models, and their variants. The important performance measures of the system obtained by queueing analysis are the part-type-wise values of the mean lead time, mean inventory level, and the mean machine utilisation