Specification of Resource Allocation Systems; a STEP towards a unified framework

Doktorsavhandling, 2005

In recent years growing demands on flexibility and ability to decrease time to market has made it increasingly important for engineering companies to find ways of making information exchange between product design and manufacturing systems design more efficient. A much shortened iteration cycle could be obtained if information about product design solutions could be made instantly available for engineers involved in manufacturing systems design. Due to the high costs associated with modifying and changing system implementations, the ability to model and simulate systems before they are implemented is becoming more and more essential. Consequently it is vital that the system specifications used are as clear and concise as possible. The present thesis deals with the specification of discrete event systems, especially resource allocation systems. A combination of process algebra and Petri nets is presented. This combination results in a powerful language, called process algebra Petri nets (PPN), for specifying resource allocation systems, delivering both concise and easy-to-read specifications of large complex systems. The fact that both Petri net constructs and algebra expressions can be used in order to decrease specification complexity also makes it a flexible language. A method is also presented that formally converts the PPN models into finite state automata, which means that existing formal evaluation techniques for simulation, verification, and controller synthesis can be easily applied. The presented language defines an algebra where the process operators express the same process relations that are possible in the international standard STEP-AP214. To the best of our knowledge, the PPN language constitutes a first attempt at using a formal language in order to create a tool that can automatically generate specifications according to the STEP standard. So far little has been investigated concerning the connection between information modelling and discrete event systems. The present work, however, researches this connection. The presented mapping defines the relationship between the information and the DES specification. Finally, it can be said that the introduced method guarantees that the expected information is delivered fast and without the errors potentially induced by manual handling, something which is crucial when short lead times are required. Due to the fast information exchange it also enables simulation, automatic controller synthesis and verification, to be conducted early in the development chain.