A neural controller for transit node in high speed communication networks

In order to provide multi-megabit communication services to an increasing number of users one needs efficient means of traffic control and network resource management. The neural network-based controller used in transit node of high speed networks seems to be a useful tool for assurance reliable and cost effective communications. The major aim of this paper is to study the problem of traffic control in complex heterogeneous networks as well as the usefulness of neural nets for design of a controller allowing very fast responses and adaptability to changing traffic environment.