DISTRIBUTED INTELLIGENCE IN MANUFACTURING INFORMATION SYSTEMS

Abstract

This paper describes an information system that was architected to enable front end distributed intelligence in a real manufacturing environment. By distributing intelligence to the most robust components of the system, data can be preserved at the cost of very real-time information. In order for the system as a whole to function this way, each software component was fitted with watchdog agent and granted system level access to the operating system so that it could detect and repair itself and other tasks in the event of a failure. During the repair process, the appropriate data collection processes were granted a level of autonomy that preserved data integrity. The paper is quite practical and designed as a case study rather than a theoretical research paper, although the techniques and principles applied are the result of much study.