Autonomous network QoS control by the user adaptive agent and its evaluation.

The Internet is a wide area distributed network emphasizing the importance of interoperability and scalability and takes a bottom-up approach by stacking local schemata. It is considered a dynamic network environment including nondeterministic and nonformatted characteristics. In this paper we propose a platform involving user adaptive agent with a mechanism to realize deliberative responses in order to carry out predictive learning of the environmental variations, not just reactive responses. Further, the present platform is applied to a network QoS control system. It is now possible to provide for graceful degradation while the trade-off point between the user and the network is found autonomously. © 2000 Scripta Technica, Electron Comm Jpn Pt 2, 83(4): 1–12, 2000 [ABSTRACT FROM AUTHOR]