Distributed intelligent control for a mine refrigeration system

One way to construct intelligent controllers is to use an agent-centered approach in which the agents themselves determine the global structure of the controller and the inter-agent cooperation methods. To assess this approach, the design and testing of a distributed intelligent controller for a laboratory-scale mine refrigeration plant is discussed. Underlying theoretical concepts are briefly reviewed, and experimental results are presented. Important findings are that the approach supports the use of multiple knowledge representations, the use of partial results, and dynamic structuring of the distributed controller. The results presented indicate that development for practical, supervisory-level control applications is worthy of further investigation.