Agent-based models can be applied to many situations where large groups of individuals come together to create some sort of cohesive behavior. The dynamics of such a group can vary greatly depending on how the individuals interact. I will present my past and current work on such agent-based simulations. In particular, I will
1) describe the dynamics of a a Czirok-Vicsek-type model for fish migration which has shown promising predictive results
2) derive and discuss a spin-system-type model for gang territoriality using graffiti as a territorial marker
3) describe a model which we are currently designing, together with criminologists at UCI, to explore the causes of gang violence in Los Angeles.