A Cellular Automaton in RA/CA Crime Simulation

Lin Liu, University of Cincinnati
Xuguang Wang, University of Cincinnati
John E. Eck, University of Cincinnati
Jun Liang, University of Cincinnati

ABSTRACT
Cellular automata (CA) provide a general framework for spatial simulation. Various studies have applied CA in simulating forest fire, urban sprawl, transportation, etc. It is not until recently has CA been applied to crime simulation. A CA has four elements: a cell space, a state variable, a neighborhood template and a set of transition rules.

This paper addresses main issues surrounding the design of a CA for simulating crime processes and events, with a focus on the state variable and transition rules. The overall reaction of people to crime events is represented as an abstract variable - tension, and tension is used as the state variable in the crime simulation model. The authors illustrate how the impact of a crime event can be modeled as a space-time process by implementing transition rules for spatial decay and temporal decay of the crime. This CA, tightly integrated with routine activities and agent based learning, serves as the foundation of a RA/CA crime simulation model, which has generated some plausible results for studying crime processes.

(Return to Program Resources)

Updated 05/20/2006