Enterprise integration with agent-based engineering data management
IFIP Advances in Information and Communication Technology
Information is the key factor in industry nowadays. Especially Computer Integrated Manufacturing intends to use information technology to integrate all parts of an enterprise. Product development often involves large teams of design specialists, produces a large amount of data and should consume as less time as possible. Therefore, CIM forces the application of information management technology in order to guarantee that the right data are present at the right location in the right time. Agents
... and multi-agent systems are an upcoming technology that can fi.elp to establish distributed intelligent applications for industrial environments. Agent-based applications seem well suited to the distributed character of an enterprise-wide information management. However, industry requires a data management that is robust and reliable. In this paper we want to show how database technology can be integrated with agent technology to provide a persistent, robust, and reliable information management within multi-agent systems. Collaborative product development is a situation where different designers are working together exchanging experience and knowledge for fulfilling their design task. Supporting distributed design teams with information management means to maintain large amounts of data of different engineering disciplines. All these data are strongly interrelated and form a «web of information.» The process of product development is rather evolutionary and characterised by iterations of design activities that initialise, transform and extend such a «web of (product) information.» Many restrictions like special requirements of specific production processes, material constraints, and technological aspects must be taken into account. Designers of different engineering disciplines and with specific distinct experience and knowledge are involved in this process of product development. They build a kind of «web of shared understanding» by discussing product details and making design decisions. Agent technology is a new and attractive way of implementing intelligent and distributed information systems. The concept of co-operating agents seems well suited to the implementation of information systems that can automatically reason and keep track of the various restrictions and technological aspects of a design process. It seems as if this «web of shared understanding» can also be manifested in software using agent technology. However, for the industrial use of multi-agent systems a secure and reliable data management is very important. The research project «IDEE Intelligent Data Engineering environment» is an effort to combine concepts of expert system shells and database system technology to enable the modelling and implementation of multi-agent systems. With IDEE a persistent and reliable data management can be provided for this «web of shared understanding.» IDEE demonstrates how a database system in combination with an expert system shell can build an environment that allows the development of industrial applications in an agent-oriented style. Currently we are able to implement agent-based applications in which the agents have reasoning capabilities, can be distributed within a computer network, and have access to an active realtime database system (Purimetla, 1993 , Ramamritham, 1993 , Chakravarthy, 1989 , Dittrich, 1995 . The database system is used by agents for communication and exchanging data and provides a network-wide event notification mechanism. Intelligent data engineering environment IDEE is based on the concepts of intelligent agents (Wooldridge, 1995) and supports the implementation of multi-agent systems. As agents are «intelligent autonomous software entities» that «co-operate» with each other, IDEE must support the concurrent execution of a set of agents and enable co-operation. Co-operation means that agents work together to solve a specific problem. Therefore communication is essential for multi-agent systems. Agents communicate by exchanging «speech acts» (Austin, 1962 , Cohen, 1988 , Searle, 1969 , Searle, 1985 . KQML (Knowledge Query and Manipulation existing applications. Persistence is a very important concept within the IDEE-approach. Persistence allows an IDEE-agent to permanently store ontological information, and it enables the communication of these agents. First we want to show how agent communication was implemented for IDEE using the database system as a communication facility, then we show how CLIPS-objects can be made persistent for IDEE, how persistence and communication are interrelated, and how both features enable the concept of distributed persistent objects for IDEE.