Assessing the operational analysis robustness of flexible manufacturing systems modelling

An effective tool for modelling complex production systems such as flexible manufacturing systems (FMS) is the queueing network theory. In this framework, operational analysis is a recent model that does not require any stochastic assumption and is based on a set of operational hypotheses related to directly observable properties. Applications of operational analysis to FMS performance prediction involve the assumption of homogeneous service times for all work stations. This is the most restrictive assumption of the operational approach. We present a new technique for obtaining the infimum and the maximum for throughput and utilizations when all work stations arbitrarily deviate from homogeneous service time behaviour. This paper proves that the robustness of resolvent formulae is not an absolute characteristic of the operational model.