Handbook of neural computer applications
Handbook of neural computer applications
Observations on the dynamics of a congestion control algorithm: the effects of two-way traffic
SIGCOMM '91 Proceedings of the conference on Communications architecture & protocols
Neural nets for adaptive filtering and adaptive pattern recognition
An introduction to neural and electronic networks
Flow control and discretized learning algorithms for routing in computer networks
Computer Communications
Traffic management in frame relay networks
Computer Networks and ISDN Systems
Traffic measurements on a local area computer network
Computer Communications
Hi-index | 0.00 |
An effective method to execute the traffic management of computer networks using a cascading neural-net (CNN) is proposed in this paper. The proposed CNN consists of a two-level neural model. The lower level, the back-propagation neural model, will detect whether the tested network is overloaded or not. The higher level, the counter-propagation neural model, will classify and exclude the status of congestion derived from the overload of tested network. Therefore, if the diagnostic effect gained from the lower level neural model is positive, the higher level neural model will be triggered to map one of flow exemplars to reroute the traffic of computer networks. These two-level neural model is iteratively and interchangeably executed to achieve the objective of traffic management. To validate the feasibility, the proposed CNN has been applied to a LAN environment. The experimental results demonstrate that the developed CNN can efficiently and effectively provide substantial assistance for decision making in network traffic management.