Agent support for operating and maintaining automation systems

Abstract Software agents are tools wherein individual agents present their elaborated information or are in addition cooperating for exchanging the knowledge of their specific domain before presenting it. They can do decisions automatically without intervention of human operators or they can assist the decision making of human operators. The contribution reports on solving maintenance problems where software agents assist operators. In principle agents can learn what is called machine learning, but it is emphasised that It could be done cooperatively and with the human operators. A general consideration of the structure of decision supporting systems are also discussed.