FAUTH, J.E.*; OLESON, R.; KAUP, D.J.; MALONE, L.C.; CLARKE, T.L.; University of Central Florida (UCF); UCF; UCF; UCF; UCF: Using crowd-control simulations to predict the behavior of motile organisms

Simulation is being used to develop crowd-control models to understand the collective behavior of pedestrians in the street or of people finding their way out of a room or building. The goal is to understand and modify crowd behavior to reduce fatalities during emergencies such as nightclub fires, stadium accidents, and subway or building bombings and other terrorist acts. While crowd-control models are useful for understanding and predicting human behavior, they also can be applied to other motile, multicellular organisms. For example, models can be developed to predict the behavior of animals using ecopassages, the outcome of complex interactions among competing organisms, or the spread of an invasive species. We expanded and generalized the pedestrian motion model by Helbing-Moln�ar-Farkas-Vicsek and used it to investigate a classic biological phenomenon: niche partitioning by salamanders along the streambank-forest floor ecotone. Our model contained four different classes of forces: 1) an action-reaction force between individual salamanders; 2) environmental forces acting on each individual; 3) preferred velocity of movement; and 4) goal forces that reflect movement of individuals toward preferred breeding sites and diurnal refuges. Each of these forces included both physical and social forces, the latter being an �action-at-a-distance� type force that reflects responses to visual and pheromonal cues. We evaluated model predictions by comparing substrate-choice selection of simulated salamanders with those of real salamanders interacting within experimental mesocosms. Simulations produced realistic salamander behavior, including non-overlapping territories and interspecific niche partitioning.