A distributed AI architecture for customer network control

A telecommunication network management application of distributed artificial intelligence research on cooperative distributed problem solving between dissimilar autonomous agents is described. From small personal subnets to large corporate networks, private network management and customer control of private networks ranks as an important new class of telecommunication services. Distributed network management automation systems, collections of intelligent, but dissimilar, agents that cooperate to solve common service provisioning and facility allocation problems, are proposed. Such agents form the basis of an architecture for joint private and public network management and customer network control. A research testbed, TEAM-CPS, that explores cooperative problem solving among such agents is described. Interagent cooperation in the testbed is illustrated with a simplified dialog between two agents jointly solving a facility failure problem.<>