M. Wiedijk

A Federated Cooperation Architecture for Expert Systems Involved in Layout Optimization

In shoe and handbag industry, there are many CIM activities that can at best be represented by expert systems. One example of such an activity is the optimization system that supports the cutting of regular/irregular shapes out of uniform/non-uniform base leather material. This paper focuses on and describes a layout optimization environment with five different kinds of agents, where every agent is an intelligent expert system. In order to perform the necessary tasks in the layout optimization activity, the involved agents need to cooperate, and share and exchange their information both with each other and with other agents involved in other CIM activities. In general, an agent representing a CIM subactivity can be represented either by an intelligent system, a simple file system, or an information management system. A natural architecture to represent such agents’ cooperation and information sharing is a federated/distributed network of interrelated agents. In this paper we first address and describe the peer federated architecture and then present a peer implementation of the agents involved in the layout optimization system and the architecture to support their cooperation.

Supporting the Information Management of Control Activities in a CIM Shop Floor Environment

A CIM shopfloor environment can be supported by a collection of different activities, such as design, manufacturing, marketing, etc. Every activity may consist of a number of subactivities where a subactivity can be performed and represented by several agents. The “agents” are de- scribed here as components that perform activities through executing their specific “tasks”. So far, for DCS (Distributed Control Systems), manufacturers have resolved the “information management”, “information sharing” and “control” problems, through purchasing all instruments and control equipments from the same supplier. In real applications however, both within the CAC activity (among its sub-activities) and in relations among the CAC activity and other CIM activities, the industrial processes define differently formatted data, for instance the discrete and continuous variables that need to be shared among several multivendor control equipments. Therefore, the information sharing and management and the control problems are still the difficult technological problems without an integral solution. This paper focuses on the Computer Aided Control (CAC) activity and describes some agents involved in process control and the tasks that they perform. Furthermore, an architecture is described to support the information management requirements for agents’ cooperation in a CIM shopfloor environment. This information management architecture is then applied to the definition of cooperating CAC agents.

Information Management for Intelligent Supervision of Hybrid Process Control

The main objectives in the design of Hybrid Process Intelligent Supervision System (HPISS) are the supervision of hybrid (i.e. discrete, continuous, and symbolic) process control, the structured information management among its different components, and its integration with the process control and other CIM activities. The HPISS assists the supervisor during the normal operation and in emergency states. During normal operation the supervision system supports the management of production goals and the supervision of the process control. Both contingency analysis and the support for decision making assist the supervisor in emergency situations. This paper describes the design of HPISS as a multi-agent system with the emphasis on sharing and exchange of information among the agents, using the PEER federated integration architecture. The supervision of a hydro-electrical plant is described in the paper as an example application.