Design of collaborative and reconfigurable control architecture for dispersed productive systems

In general, the control systems for conventional manufacturing systems are organized into hierarchical layers such as, enterprise control, factory control, and shop floor control. In an approach based on division in scope and breadth of the problems these levels are designed to act with relative autonomy. However, currently manufacturing systems must necessarily be reconfigurable to assure features of adaptability to unpredictable changes, agility, flexibility and efficiency, i.e., the parts of the reconfigurable manufacturing system (RMS) must collaborate each other. A RMS must combine a global vision with a relatively agile reaction to unexpected variations of the plant. On the other hand, collaborative activities in RMS can be treated as services and therefore, service oriented architecture (SOA) can be effective for integration of the control layers. Besides, in a global market, the RMS must also consider a reconfiguration of subsystems that can be geographically dispersed and that are processing information, physical transformations and material handling. In this context, the integration of multi-agent system and holonic system techniques named as holonic control system (HCS) has been considered a promissory tendency for the control system of the RMS. Therefore, this paper introduces a procedure to design a service oriented active holonic control system (SOAHCS) and also presents an application example for a benchmark dispersed productive system. SOAHCS combines bottom–up and top–down approaches using Petri net technique, and its extensions to represent the structure and control operations under faults occurrence, and explores the combination of HCS and SOA techniques, such as autonomy, reactivity, proactivity, cooperation, learning, reuse, interoperability and