Game Theory in Crime Prevention
Imagine a world where crimes could be predicted and stopped before they can ever occur. Typically, this would need someone with precognition abilities to be achieved effectively. However, there is a team of researchers at Vanderbilt are working together to make this goal achievable through data, computing, and analysis. Essentially, the researchers are applying game theory to create a mathematical model of expected crime behavior so that the police can make the appropriate decisions to best prevent crime.
There are three main methods that police already use to quickly respond or even rapidly respond to crimes. The most popular method is perhaps to use crime data to determine active spots for criminal activity so that the police can proactively patrol these areas. A second approach is terran risk-based modeling, which uses terrain layers (such as bars, stores, shops) to help pinpoint areas of highest crime incidences. A third method, which is developed relatively recently, is to use a mathematical crime diffusion model based on environmental factors and past crime locations to predict crime incidences.
Vorobeychik, an assistant professor in computer science at Vanderbilt, stated that the researchers are incorporated game theory methods that are largely successful for several locations including canine patrols at the Los Angeles airport and US Coast Guard patrol scheduling at New York harbor. Though it is extremely unlikely that crimes can be pinpointed precisely, Vorobeychiik proposed that a probability distribution of crime in time and space to assess the best expected spots for police to look out for. This distribution will integrate all of the three main methods, which will be experimented in various degrees to determine the most effective model for the police to adopt.
http://engineering.vanderbilt.edu/news/2014/game-theory-can-help-predict-crime-before-it-occurs/