Decision Making in Distributed Objects Using Radio Frequency Identification

This paper explores the architectural and management challenges related to the structure of a distributed production control application and its integration with the physical hardware comprising the production line. We propose that management control, risk management and information management can be enhanced through an integrated, agent-based systems architecture that attaches data processing and storage capabilities to physical objects. The objective of our research is to extend prior research and development in integrated systems architectures by using radio frequency identification (RFID) technology. RFID technology will allow much of the coordination and decision making to take place directly in the RFID tags, rather than in digital surrogates within the software application. This projects the digital representation directly into the physical world, minimizing the intermediary and somewhat centralized software-based system, and creating a more natural representation of a physical, object-based system. INTRODUCTION AND MOTIVATION Advances in sophisticated data storage and wireless communications technologies, such as radio frequency identification (RFID) devices, provide opportunities to incorporate object-oriented (O-O) processing and design techniques within the physical world. (Nicolai et al., 2005) This is a natural evolution of the object-oriented software paradigm, given that intelligent software objects, or agents, are intended to mimic the structure and behavior of entities in the real world, thereby allowing a more ‘natural’ model of a complex engineering, biological or social system. (Jennings, 2001) Likewise, the application of O-O design and control techniques within an RFID-enabled system can lead to improved flexibility, agility and robustness in a variety of physical environments, including transportation, health care, and inventory control. (Nath et al., 2006) This paper proposes that management control, risk management and decision making within a production environment can be enhanced through an integrated systems architecture that incorporates decision making capabilities within physical objects. We focus specifically on the design and control issues of an object-oriented distributed application involving physical objects, such as machines and materials, casting them as intelligent agents in the context of a potential implementation environment such as RFID. The O-O implementation intentionally leads to a distributed, loosely coupled information system infrastructure, where processing, data storage, and control is assigned, via RFID tags and technology, to physical objects making decisions within the production domain. Assigning distributed responsibilities to individual objects raises important questions about the ability of individual RFID tags to act contingently in a dynamic physical environment. That is, to what extent can a tag plan or re-plan based on environmental parameters such as its location, date/time, and proximity to other tags or readers? Furthermore, how is this contingency planning controlled by the overall system? In many respects, such control issues are addressed within the context of an object-oriented design. An objector agent-oriented approach is motivated by a disconnect between the physical operation of a system, which is inherently decentralized, and its management and control structure, which tends to be centralized. Consider, for example, a manufacturing environment in which the production line is dynamic and produces a number of different items. The actual execution of the system – including operations, queues, workstation setups, and movement of items – is physically decentralized. By contrast, the associated manufacturing and control system is managed through a