Meeting Abstract

100-3   10:30 - 10:45  Uumarrty: Agent Based Simulation Model of Predator Prey Interactions in a Game Theoretical Framework Remington, M*; Higham, T; Clark, R; Sukumaran, J; San Diego State University; University of California, Riverside; San Diego State University; San Diego State University michaelremington2@gmail.com

Uumarrty is a Python software package used to conduct experiments on how ecological and climatic changes affect the microhabitat preferences of a predator and a prey agent. Our model is a two-dimensional virtual landscape composed of individual cells classified as either an open microhabitat or a bush microhabitat. A set of agents move on this landscape through semi Brownian motion and based on encoded microhabitat preferences. Agents look to survive while also optimizing payoffs through hunting or foraging strategies throughout their generation time. In this talk, we utilize Uumarrty to analyze the predator-prey dynamics between kangaroo rats (Dipomys sp.) and rattlesnakes (Crotalus sp.). Research has shown microhabitat usage is an essential component in the historical competition between the rattlesnake and kangaroo rats. Previous deterministic game theoretic models have done a good job of analyzing preferences under a wide variety of conditions and establishing the evolutionary stable strategies of these two agents. However, our model allows us to conduct virtual experiments on these populations in a novel manner and shed new light on this system that would be hard to accomplish in field experiments or through traditional modelling methods. Analyzing these two agents competing in a game theoretic framework while allowing strategies to mutate throughout generations provides time series data which gives us a dynamic view of the performance of certain strategies. This data so far allows us to visualize the cyclical fluctuations of microhabitat preferences between these two competitors. Post simulations we will be able to view the Nash Equilibrium set of strategies that optimize the payoffs for both agents under differing experimental designs.