Myopia of Service Oriented Manufacturing Systems: Benefits of Data Centralization with a Discrete-Event Observer

Abstract

Service orientation paradigm is particularly well adapted to distributed manufacturing systems. The difficulty of such systems’ production activity control deals with knowledge management. Indeed, the knowledge is distributed among each entity which is able to create, modify or communicate with other entities. An entity cannot have a full up-to-date access to all the system’s data. At shop floor level, a convenient way to implement service-oriented manufacturing systems is to rely on the paradigm of Holonic Manufacturing Systems. The chapter introduces the possibility of specializing a resource holon in order to gather data from the whole holarchy and make these data available to any holon for a decision making. This holon is thus playing the role of a discrete-event observer. Upon positioning the service-oriented architectures, the HMS reference architecture PROSA is described, especially in terms of decision making. After decisions were defined, the problem of online decision making in a HMS is described, and an implementing solution for the observer and the forecasting tools in the architecture is exposed. Finally, two applications are presented, based on an industrial job-shop.