In-node fast optimal packet routing algorithm in Manhattan street network using cellular neural network

An in-node fast optimal packet routing algorithm in Manhattan street network using cellular neural network (CNN) is described. The proposed algorithm generates not only the possible transition matrix (governing the switching in the node) but also the optimal transition matrix, taking into account the traffic density in the vicinity of the node and the sum of packet priorities. In this way, the overall throughput rate was improved, compared to the shortest path algorithm. This result is proved by computer simulations. The inherent CNN's feature of parallel analog processing permits on-line processing without any additional control logic circuitry.