An object-oriented event-driven individual-based spatially-explicit simulation environment for the construction of arbitrary food webs

Abstract

Object oriented programming and event driven dynamics provide a methodology for constructing individual-based simulations with flexibility and clarity by allowing a simulation environment to closely mimic its real world counterparts. A simulation environment, Assisi Predator Prey, was built to investigate the effectiveness of using such techniques to model food webs at the individual level. Models of herbivory, competition and predator mediated coexistence were constructed and their dynamics compared with historically expected behavior. The simulations were very sensitive to model parameters and could be made to display a wide variety of dynamics. Object oriented methods allowed a close connection between model components and their real world counterparts. Although the simulation environment produced complex dynamics, simulation structure remained relatively simple. Event driven dynamics allowed time dynamics to more closely match simulation object behaviors than incrementally based dynamics but at the cost of less knowledge of the flow of simulation time. The results support the hypothesis that system behavior at the scale of populations and communities are emergent properties of interactions at the individual level.