Crime Simulation Using GIS and Artificial Intelligent Agents

Abstract

This chapter presents an innovative agent-based model for crime simulation. The model is built on the integration of geographic information systems (GIS) and artificial intelligence (AI) technologies. An AI algorithm (reinforcement learning) is applied in designing mobile agents that can find their ways on a street network. The multi-agent system is implemented as a Windows desktop program and then loosely coupled with ESRI ArcGIS. The model allows users to create artificial societies which consist of offender agents, target agents, and crime places for crime pattern simulation purposes. This is a theory-driven system, in which the processes that generate crime events are explicitly modeled. The simulated crime patterns are shown to have similar properties as seen in reported crime patterns.