A Mean-Field Theory of Cellular Automata Model for Distributed Packet Networks

A Mean-Field theory is presented and applied to a Cellular Automata model of distributed packet-switched networks. It is proved that, under a certain set of assumptions, the critical input traffic is inversely proportional to the free packet delay of the model. The applicability of Mean-Field theory in queue length estimation is also investigated. Results of theoretical derivations are compared with simulation samples to demonstrate the availability of the Mean-Field approach.