Intelligent management of computer networks: the GIR proposal

The paper presents a system proposed for the intelligent management of computer networks. The system is based on the use of artificial intelligence techniques-data mining, expert systems and multi-agent systems. The work is based on the following lines of action: (a) the use of distributed agents for the intelligent search of information in the network, supplying it in a more abstract way, adapted to the decision-making task; (b) the use of machine learning and data-mining techniques that, starting from the log files which register previous problems and their solution, allow the use of experience thus obtained in the solution of new problems; and (c) the use of heuristics and conduction rules supplied by experts, through a decision support system, as an advisor element to network operators. The paper presents a discussion about the techniques employed along these three research lines and the results already obtained