A distributed AI architecture for customer network control

Cooperative multiagent systems: a personal view of the state of the art

“…The need for a multiagent approach can also come from applications in which agents represent the interests of different organizational entities (e.g., electronic commerce [40] and enterprise integration [2]). Other emerging uses of multiagent systems are in layered systems architectures in which agents at different layers need to coordinate their decisions (e.g., to achieve appropriate configurations of resources and computational processing [53]), and in the design of survivable systems in which agents dynamically reorganize to respond to changes in resource availability, software and hardware malfunction, and intrusions. In general, multiagent systems provide a framework in which both the inherent distribution of processing and information in an application and the complexities that come from issues of scale can be handled in a natural way.…”

“…This lack of a perfect fit often leads to a situation where there may be insufficient local information or resources for an agent to completely or accurately solve its assigned subproblems through its own processing. Further, resource contention issues in multiagent systems do not entail simply reasoning about exclusive access to a shared resource, but may involve more subtle issues such as what percentage of a resource's capacity an agent will use (e.g., a communication channel), the creation of shared agent plans so that the use of a scarce resource can satisfy multiple objectives [52], or the reconfiguration of resources to better meet the competing needs of agents [53].…”

Active Network Management via Agent Technology

“…It can be effectively used to decentralize network management activities [5]. In particular the characteristic mobility, co-operation and autonomy of the agents emphasize the advantages of AOT in distributed management applications.…”