Spatio-temporal Organization of a Cellular Automaton Model for Computer Network

ZL Ren,ZD Deng,ZQ Sun
DOI: https://doi.org/10.1016/s0010-4655(02)00272-2
IF: 4.717
2002-01-01
Computer Physics Communications
Abstract:It was unveiled by Ren et al. [Comput. Phys. Comm. (2001)] that congestion transition emerges in cellular automaton models for computer network and this NaSch network model with Q=1 has similar behaviours as the NaSch traffic model with maximum velocity vmax=1. For these two NaSch models, the main difference lies in a node cell contained in the NaSch network model. In this paper, we will focus on our further investigation on spatio-temporal organization of the NaSch network model. More interesting phenomena of phase transition are discovered. Firstly, fundamental diagram illustrates that when Q>1 for the NaSch network model it is significantly different from its counterpart, i.e. the NaSch traffic model in a road traffic system. The addition of a node cell, which is allowed to have more than one packets, will lead to generating a new phase. Secondly, in order to characterize phase transition occurred in the NaSch network model, an order parameter is presented with the use of the time average density of nearest-neighbor pairs m. The computational results obtained show that criticality will disappear in a strict sense if noise exists. Finally, two other numerical features, i.e. spatial correlation functions G(r) and relaxation times τ, are analyzed so as to deeply describe behaviours near critical points.
What problem does this paper attempt to address?