Mobile agents for data analysis in industrial automation systems

In modern power systems and other industrial concerns, a large quantity of data is produced by condition monitoring and control systems. This data is used for a variety of purposes, and is often distributed over a number of sites. This paper describes part of a prototype system which combines mobile agents with a Foundation for Intelligent Physical Agents (FIPA)-based multi-agent system to provide remote analysis and reporting functions for data stored in power substation databases. A model of mobile agent performance in this application is derived, and a number of experimental results regarding the relative performance of mobile agents and static client-server agents are presented. These results suggest that in certain circumstances mobile agents have the potential to greatly increase the performance of this application. A comparison between using wrapper agents and direct database access for data retrieval is also performed.