Learning in the recurrent random neural network
Neural Computation
Explicit allocation of best-effort packet delivery service
IEEE/ACM Transactions on Networking (TON)
Measurement and performance of a cognitive packet network
Computer Networks: The International Journal of Computer and Telecommunications Networking - Special issue on networking middleware: selected papers from the TERENA networking conference 2001
Impact of TCP-like congestion control on the throughput of multicast groups
IEEE/ACM Transactions on Networking (TON)
Performance Evaluation
Learning to Predict by the Methods of Temporal Differences
Machine Learning
Networking with Cognitive Packets
ICANN '02 Proceedings of the International Conference on Artificial Neural Networks
Cognitive Packet Networks: QoS and Performance
MASCOTS '02 Proceedings of the 10th IEEE International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunications Systems
Block loss reduction in ATM networks
Computer Communications
QoS routing for anycast communications: motivation and an architecture for DiffServ networks
IEEE Communications Magazine
QoS routing granularity in MPLS networks
IEEE Communications Magazine
Restorable dynamic quality of service routing
IEEE Communications Magazine
Distributed quality-of-service routing in ad hoc networks
IEEE Journal on Selected Areas in Communications
Learning in the multiple class random neural network
IEEE Transactions on Neural Networks
Hi-index | 0.00 |
We show how "self-awareness", through on-line self-monitoring and measurement, coupled with intelligent adaptive behaviour in response to observed data, can be used to offer quality of service to network users. We first describe the general principles which govern our design, and briefly describe the experimental packet network system we have built in which users are allowed to specify their QoS objectives. The network uses on-line adaptive traffic routing to try to meet the users' QoS requests. Cognitive or smart packets are used for self-observation, and reinforcement learning with neural networks is implemented at network nodes to seek new paths and deduce improved paths from existing routes. First we show how the network is able to discover routes, beginning with an "empty state" and starting from a random search. Secondly we show how our network can intelligently direct traffic through the Internet to optimize web traffic for a user by offering the best quality of service through different Internet Service Providers.