Funded by the US National Science Foundation
 
 

A geographic profile of a Los Angeles offender who committed a series of more than 33 crimes. The profile provides an means of estimating the most likely "anchor point" for an offender such as his home location. The algorithm and software to produce this profile were developed by undergraduate math and computer science students Derek Lietz, Anca Dragan, Laney Kuenzel and Xinyuan Xu for the Los Angeles Police Department. The work was conducted as part of the UCLA Institute of Pure and Applied Mathematics RIPS program, which brings together promising young students from around the world with industry to tackle real-world problems.
 

Simulating gang violence using an excited Hawkes Process. The top two pannels show real shootings between rival gangs in Hollenbeck Division of Los Angeles. The bottom two panels show a simulated series of shootings. Green points represent the timing of each between-gang shooting. Blue curves represent the time-course of the rate function describing a Hawkes Proess. This work was completed as part of a UCLA Applied Math REU project involving undergraduates Mike Egesdal, Chris Fathauer, Kym Louie and Jeremy Neuman in collaboration the LAPD.
 



Two-dimensional discrete simulation of burglary hotspots in an 18 x 18 km area of the San Fernando Valley, Los Angeles. Burglary Hotspot field (left) and spatial distribution of crimes (right). Model parameters are calibrated against housing density and adjusted to yeield the best fit for Ripley's K statistic when comparing the spatial distributionm of true and simualted crime events. Simulations by Martin Short (UCLA Math).

 



Two-dimensional discrete simulation of burglary hotspots patrolled using different polcing strategies. Left: "Hot spot" policing where police (blue) gravitate towards the highest density crime areas is seen to quickly reduce overal crime with a simple deterent effect. Right: Random patrol where police follow a simple random walk. Deterence does not reduce crime in the simple random walk case. Burglars are shown in green. Simulations are part of the PhD thesis work of UCLA Math student Paul Jones.

Two-dimensional continuum simulation of burglary hotspots. The model is based on an advection-reaction-diffusion equation. Burglary rates are represented by color: green corresponds to the equilibrium rate; purple correspons to no crime; red corresponds to two times the equilibrium rate. The simulations were developed by Martin Short, a UCLA Postdoc on the UC MaSC Project.



hits since 06-22-06