Neural network-based routing in an intelligent computer communication network

A neural network-based traffic routing algorithm, to be implemented in an intelligent computer communication network, is presented. It simultaneously takes into account the shortest path and channel capacity and determines the necessary connections and weights of the neural network, thus providing an optimum transmission path from source node to destination node. The algorithm is based on the Hopfield type of neural network approach, and can be considered as a massively parallel distributed processing system with the ability to improve performance through dynamic learning. The power of neural network algorithms to solve routing problems and provide dynamic response to traffic conditions measured throughout the network system is investigated. No congestion occurs in this network because of the changes in the statistics of the input traffic load. A computer simulation result for a 10-node network shows the efficiency of using the neural network-based algorithm