The Simulation of the Journey to Residential Burglary

This chapter presents an innovative approach to the study of the journey to residential burglary. We demonstrate a simulation model that is built upon the integration of cellular automaton (CA) and multi-agent system (MAS). The model utilizes both social disorganization (SD) and routine activity (RA) theories to predict locations of residential burglary targets. The model simulates an offender as an intelligent agent of MAS that interacts with the target and place automata of CA. The likelihood of a residential burglary is modeled as a function of offender’s motivation, target desirability and place guardianship, which in turn are determined by the offender’s individual characteristics often used by RA and the target and place’s neighborhood properties frequently utilized in SD. The model was initialized and parameterized using “real” crime data from Dallas, Texas Police Department. Results under two different weighting scenarios were obtained and compared with the actual distribution of offense locations, revealing the flexibility of model in its capability to assessing the influence of the two theories in burglary crime simulation. In closing we propose possible modifications that could be made to the model in the future.

Mastermind

Pattern and routine activities theories suggest that through a combination of decisions and innate understandings of environmental cues, likely offenders are able to separate good criminal opportunities from bad risks. The nature of this process is highly structured and allows researchers to anticipate likely concentrations for a variety of regular, daily activities, including criminal offences. This chapter sets out to model and test these theoretical principles. Mastermind represents an interdisciplinary research project in computational criminology jointly managed by ICURS and the Software Technology Lab at Simon Fraser University. Using the abstract state machine (ASM) formalism in combination with a multiagent based modeling paradigm, we devise a formal framework for semantic modeling and systematic integration of the theories for crime analysis and understanding crime patterns. We focus on crime in urban areas and model spatial and temporal aspects of crime potentially involving multiple offenders and multiple targets. Mastermind is used in a hypothetical analysis of motor vehicle theft.

Core Models for State-of-the-Art Microscopic Traffic Simulation and Kay Elements for Applications

This chapter summarizes fundamental models for microscopic simulation (such as vehicle generation model and car-following model) and other critical models (such as lane-choice model, lane-changing model, and route-choice model). Most of the critical models introduced in this chapter reflect the latest research results by the author. The primary purpose of this chapter is to provide fundamentals for better understanding of the travel behaviors that are modeled for traffic simulations. To facilitate the applications of traffic simulation models, several key elements for applying state-of-the-art computer traffic simulation tools are summarized. They include the procedure for building models, model calibration and validation. Further more, techniques for collecting vehicle trajectory data, critical elements used for model calibration and validation, are also introduced.

Modelling Pedestrian Movement to Measure On-Street Crime Risk

This chapter presents results for the first large-scale analysis of street crime rates that utilizes accurate on-street pedestrian population estimates. Pedestrian counts were generated at the street segment level for an area in central London (UK) using a modeling process that utilized key indicators of pedestrian movement and sample observations. Geocoded street crime positioned on street segments then allowed for street crime rates to be calculated for the entire central London study area’s street network. These street crime rate measures were then compared against street crime volume patterns (e.g., hotspot maps of street crime density) and street crime rate statistics and maps that were generated from using the residential population as the denominator. The research demonstrates the utility of pedestrian modeling for generating better and more realistic measures for street crime rates, suggesting that if the residential population is used as a denominator for local level street crime analysis it may only misinform and mislead the interpretation and understanding of on- to pedestrians. The research also highlights the importance of crime rate analysis for understanding and explaining crime patterns, and suggests that with accurate analysis of crime rates, policing, and crime prevention initiatives can be improved.

Realistic Spatial Backcloth is not that Important in Agent Based Simulation Research

This chapter considers whether it is worthwhile and useful to enrich agent based spatial simulation studies in criminology with a real geographical background, such as the map of a real city? Using modern GIS tools, such an enterprise is in principle quite feasible, but we argue that in many cases this course is not only not producing more interesting results, but in fact may well be detrimental for the real reason of doing criminal simulation studies, which is understanding the underlying rules. The argument is first outlined in general, and then illustrated in the context of a given example of the ThESE perceptual deterrence simulation model (Van Baal, 2004), a model that actually is using a simple checkerboard as its spatial backcloth.

Using Varieties of Simulation Modeling for Criminal Justice System Analysis

Criminal justice systems are complex. They are composed of several major subsystems, including the police, courts, and corrections, which are in turn composed of many minor subsystems. Predicting the response of a criminal justice system to change is often difficult. Mathematical modeling and computer simulation can serve as powerful tools for understanding and anticipating the behavior of a criminal justice system when something does change. The focus of this chapter is on three different approaches to modeling and simulating criminal justice systems: process modeling, discrete event simulation, and system dynamics. Recent advances in these modeling techniques combined with recent large increases in computing power make it an ideal time to explore their application to criminal justice systems. This chapter reviews these three approaches to modeling and simulation and presents examples of their application to the British Columbia criminal justice system in order to highlight their usefulness in exploring different types of “what-if” scenarios and policy proposals.

The Need for Systematic Replication and Tests of Validity in Simulation

This chapter outlines how simulation methods might be used to make valid causal inferences in the social sciences, specifically the study of crime. We argue that significant threats to validity exist for simulation studies and that, if researchers do not actively take measures to minimize these, much of the promise of simulation will not come to pass. Further, we nominate replication as a general method to facilitate the generation of valid findings. It is anticipated, with the burgeoning interest in simulation methods in criminology, that simulation studies will be published in sufficient detail that allows researcher scrutiny and replication, with a view to developing a cumulative body of scientific knowledge.

Simulating Crime Against Properties Using Swarm Intelligence and Social Networks

Experience in the domain of criminology has shown that the spatial distribution of some types of crimes in urban centers follows Zipf’s Law in which most of the crime events are concentrated in a few places while other places have few crimes. Moreover, the temporal distribution of these crime events follows an exponential law. In order to reproduce and better understand the nuances of such crime distribution profile, we introduce in this chapter a novel multi-agent-based crime simulation model that is directly inspired by the swarm intelligence paradigm. In this model, criminals are regarded as agents endowed with the capability to pursue self-organizing behavior by considering their individual (local) activities as well as the influence of other criminals pertaining to their social networks. Through controlled experiments with the simulation model, we could indeed observe that self-organization phenomena (i.e., criminal behavior toward crime) emerge as the result of both individual and social learning factors. As expected, our experiments reveal that the spatial distribution of crime occurrences achieved with the simulation model provides a good approximation of the real-crime data distribution. A detailed analysis of the social aspect is also conducted here as this factor is shown to be instrumental for the accurate reproduction of the spatial pattern of crime occurrences.

Crime Simulation Using GIS and Artificial Intelligent Agents

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.

Simulating Urban Dynamics Using Cellular Automata

This chapter introduces the concepts of cellular automata (CA) which have been increasingly used for simulating urban dynamics. Simulation and prediction of urban evolution can provide the useful inputs to crime models. However, calibration of urban cellular automata is crucial for simulating realistic cities. Simulation of multiple land use changes using CA is difficult because numerous spatial variables and parameters have to be utilized. The incorporation of neural networks with CA can alleviate the calibration problems. This chapter illustrates how complex land use dynamics can be simulated by the integration of CA and neural networks.