It has long been recognized that crime tends to cluster in time and space, forming so-called crime hot spots separated by areas where there is little or no crime. Advances in digital mapping technologies over the past decade have dramatically improved our ability to recognize and also quantify some of the spatial properties of crime hot spots. The dynamic aspects of crime hot spot formation, persistence and dissipation, however, are poorly understood.
The purpose of the IPAM short program conference Crime Hotspots: Behavioral, Computational and Mathematical Models is to bring together researchers studying the micro-scale behavioral and environmental bases of criminal activities with those who have approached the emergence of crime pattern formation, or similar problems in other domains, both computationally and mathematically. The intent is for the workshop participants to learn about crime pattern formation for a variety of different perspectives, to stimulate novel approaches to the study of crime and to provide an opportunity to forge new research collaborations.
Crime hot spots are defined as geographical areas with clusters of criminal offenses occurring within a specified interval of time. Hot spots may consist of clusters of property crimes such as burglaries or auto thefts, or violent crimes such as homicides, which occur on time scales ranging from hours to months. Law enforcement strategies are increasingly aimed at quickly identifying and targeting hot spots as a primary means of fighting crime. However, many fundamental questions remain unanswered concerning the generation of crime hot spots, how they should be measured and interpreted, and how hot spots might be used for predicting future distributions of criminal offenses.
This conference will bring together leading criminologist, mathematicians and computer scientists for the purpose of discussing the behavioral basis of criminal activities and exploring mathematical and computational approaches to modeling crime hot spots. A great deal is known about the micro-scale behaviors of offenders and victims as well as the environment attributes that tend to either create or restrain criminal opportunities. With a few exceptions, however, research in these domains has proceeded with only limited connection to recent developments in computational and mathematical approaches to studying emergent pattern formation. The biological sciences, by contrast, have embraced broadly the idea that simple deterministic and stochastic processes, operating at local scales, may lead to incredibly rich pattern formation at higher scales. Recognition and analysis of self-organization in biological systems has had major consequences for understanding the dynamics of ecosystems, the causes of biodiversity and, importantly, the local and global processes that may interfere with such complex systems, leading to dramatic system changes.
Part of the motivation for this workshop derives from advances in agent-based or multi-agent computational modeling and GIS crime mapping. Such computational tools provide scientists the opportunity to model offender behavior at a low-level, consistent with empirical observations, explore how collections of offenders interact with their environments and assess whether such interactions lead to the generation of crime hot spots. Formal mathematical approaches are necessary for grounding computational approaches and offer tremendous potential for developing additional insights into the nature of crime hotspots.
The five-day program from January 29-February 2, 2007 will involve presentations and discussions covering several topical areas:
Kate Bowers (University College London)
P. Jeffrey Brantingham, Chair (University of California, Los Angeles (UCLA))
Lincoln Chayes (University of California, Los Angeles (UCLA))
George Rengert (Temple University)
George Tita (University of California, Irvine (UCI))