Information flow in a dynamic weighted Erdös-Rényi network with different noise-enhanced node capacities

Abstract

We simulate the information flow within a dynamic weighted random network which will serve as baseline in the study of how information flows in a real-world system that evolves through time. The random network studied by Erdös and Rényi was used to generate the model where the edges serve as channel for transmission of information and nodes serve as capacity for storing and passing information. Nodes with same capacity and nodes with varying capacity corresponds to the homogeneous and inhomogeneous case, respectively. The information flow rate, the node capacities and the dynamics of the network is determined of its significance on the information flow within the network. Linear relations on the node capacities and information flow within the network were observed in the case of the absence of noise. In the homogeneous case, best fit lines of y = 15.99x + 1.09 with R² = 0.9971 and y = 17.34x - 2.99 with R² = 0.9894 were obtained for information flow rates κ= 0.1 and κ= 1.0, respectively. In the inhomogeneous case, the best fit lines obtained were y = 6.36x + 1.26 with R² = 0.9512 and y = 6.43x + 0.08 with R² = 0.9846 for κ= 0.1 and κ= 1.0, respectively. Similar trends were observed on some cases of noise-enhanced node capacities.