New prospects of network-based urban cellular automata

Abstract

Cellular automata (CA) build on complexity theory, focus on bottom-up processes that lead to global forms, highlight non-equilibrium or disruptive events, and complement to GIS and remote sensing techniques. Therefore, urban cellular automata (i.e., the applications of CA in urban modeling and simulation) became popular since the early 1990s. However, the CA momentum gradually subsided among urban modelers a decade later due to the criticism on the simplicity and rigidity of urban CA. Michael Batty in his seminal work, The New Science of Cities, interprets cities not simply as places but as systems of networks and flows, which can be examined in better ways by using big data and emerging computational techniques. This book sets a new benchmark for urban CA modeling, which is shedding new lights to dynamic urban modeling. Inspired by the recommendations in Batty’s book, a new network-based global urban CA framework is developed in this paper. According to the science of design and planning, this paper describes the global urban CA framework at the micro-, meso-, macro-and global-scales and examines the dynamic flows (i.e. interactions) over the urban networks at the four scales by using terminologies that are often seen in the urban CA modeling literature. In addition, the paper will analyze how big data analytics affect computational implementations of the networked global urban cellular automata, including (1) multi-scales of networked urban CA spaces that lead to modifiable areal unit problems; (2) boundary discords of CA spaces that cause areal unit inconsistency problem; and (3) incomplete data that prevent from a full implementation of the global urban CA framework.