Order-preserving models for discrete event systems: Theory and applications

Flexible manufacturing systems (FMS) are characterized by the processing of different product types on the same manufacturing systems. In particular, it is possible that the paths of different product types through an FMS overlap and different product types share the same production components such as machines or conveyor belts. That is, when designing controllers for FMS, it is required to keep track of products traveling through the FMS in order to process products correctly. In particular, it is important that the sequential order of different products is captures by dynamic models of FMS. In this context, the modeling formalism of discrete event systems (DES) is suitable since it allows capturing the sequential behavior of FMS. Accordingly, this thesis develops a new modeling technique for the supervisory control of FMS in the framework of DES. In particular, the thesis considers the general case of an FMS, where different product types can share production components and production components can hold multiple products. It is first pointed out that a suitable model for such production component needs to keep track of the product type and the order of products entering and leaving production components. Then, order-preserving languages are introduced as a new model for FMS. Several important properties of such order-preserving languages are formally proved and their benefit for modeling FMS is discussed. In addition, a general method for algorithmically constructing the required order-preserving models using finite state automata is proposed. The practicability of the developed method is demonstrated by several application examples.