Distributed fuzzy agents: a framework for intelligent network monitoring

We present a distributed fuzzy agent architecture with applications to network management and control. Each agent is designed to mimic human problem solving abilities in a network environment with monitoring and control objectives. In this fuzzy agent framework, each agent is equipped with a fuzzy micro-expert system, consisting of a fuzzy knowledge-base and an inference engine. The knowledge-base is a collection of fuzzy rules derived from an expert and represented as fuzzy conditional statements. Each agent is configured to react to precepts from its local environment and also to commands decoded in messages received from other fuzzy agents. The agents´ actions have local or external behaviors. Local behaviors are mainly due to environmental sensors while external behaviors consist of collaborative control messages forwarded to other fuzzy agents