Anthropology - Presentations

Permanent link for this collectionhttps://hdl.handle.net/2022/19577

Browse

Recent Submissions

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Agent Based Modeling (ABM) experiments help evaluate hypotheses about niche separation in early Pleistocene hominid species in East Africa
    (2008) Long, Bryon; Griffith, Cameron; Sept, Jeanne
    Agent Based Models (ABM) provide computational tools with which to simulate parameters of multivariate behavioral systems and analyze with GIS the emergent patterns of agent behavior generated by different sets of experimental conditions. We have created an agent based model HOMINIDS (Hungry Omnivores Moving, Interacting, and Nesting in Independent Decision‐making Simulations), in which we define simple behavioral rules for agents who forage in simulated landscapes. The experimental landscapes are spatially explicit, dynamic, and derived from field vegetation survey data on seasonal abundance and distribution of different plant foods collected in modern analogs of Plio‐Pleistocene habitats from two different semi‐arid riparian settings. Using the original plant food field data as a baseline, we experimentally manipulate the edible productivity and abundance of different types of key plant foods, such as different species with edible roots, to evaluate the impact of plant food availability in space and time on the foraging patterns of two different hominid agents. In addition to plant foods, we can run the model with a variety of stochastically generated carcass‐feeding opportunities. Here we report statistical and graphic GIS results of repeated, single‐year experiments. The hominid agents, Australopithecus boisei and Homo ergaster, are modeled with simple attributes that influence their foraging and nesting behavior. These include energetic requirements, gut capacity, nesting rules, travel and search rules, and different cost/benefit rankings for foods based on chewing / technological abilities, and simple cooperative opportunities. By keeping the parameters of agent behavior relatively simple (e.g., no demographic or cognitive realism), we are able to evaluate the sensitivity of resulting subsistence patterns to the different behavioral settings used in the simulations. Our results show both similarities and significant differences in land‐use behavior between the two types of hominid agents which have implications for the selective advantages and archaeological visibility of hominid subsistence behaviors.