Agent Based Fuzzy ARTMAP Neural Network for Classifying the Power Plant Performance

In this paper, we present a fuzzy artmap neural network model on a power station in al-daura refinery for the multiagent process as a classifying system to improve the process real-time performance. the proposed model is a combination of the adaptive resonance theory (art) neural network and fuzzy logic control, a supervised model having high on-line classifying accuracy learning mechanism with superior performance. the model has been applied for each agent autonomously according to agent's behaviour and standard level (s.l.) control. results have shown that the fuzzy artmap neural network is able to precisely learn to classify the data fusion from the multi-agent process to three classes: class (s) when the data fusion are within the (s.l.), class (h) when the data fusion are higher than the (s.l.) and class (l) when the data fusion are under the (s.l.). also, the fuzzy artmap is able to learn the rules and the parameters accurately with low cost, high performance and less effort.