Modeling of Diffusion through a Network: A New Approach using Cellular Automata and Network Science Techniques

Abstract

Abstract : This paper investigates the sensitivity of current models that describe diffusion or the spread of infection to changes in topology and/or initial conditions. We construct a computer simulation based on cellular automata that allows for customization of network and infection scenarios. Using the simulation we run several batch runs with varied circumstances to assess the impact of small changes in topology and initial conditions upon diffusion through a network. We also build Excel models to predict long term infection and spread rates, based upon insights from the computer-based cellular automata simulation. The goal is to highlight weaknesses in current models and present alternative models that may be more realistic. Using network science and computer science techniques, we present another option for modelers to use when forming their own spread and diffusion models.