[Swarmfest2008] Abstract for oral presentation

[Athena Aktipis]

To whom it may concern,
   Below (and in the enclosed attachment), please find an abstract of my proposed oral presentation for SwarmFest2008.

Name: C. Athena Aktipis

Affiliation: University of Pennsylvania

Address: 2122 Fitzwater St
Philadelphia, PA 19146

Type: Oral presentation

A SIMPLE model for the evolution of movement and cooperation:
Social dilemmas emerge from interactions with a shared environment

Models of the evolution of cooperation typically make use of social dilemma paradigms such as prisoner's dilemmas and public goods games, modeling situations in which there is conflict between individually optimal behaviors and socially optimal ones.  Analytical models abstract away the environment in favor of payoff matrices that define the results of various combinations of behavior of the interacting individuals, obscuring the importance of interactions with a shared environment.  In the natural world, behavior that either improves or degrades the quality of the local environment can affect those who share that environment, influencing the fitness and/or behavior of nearby individuals.  In the present simulations, agents can affect one another through interaction with a shared environment.  The inclusion of these individual-environment interactions in the SIMPLE (Simulation of agent Interactions via Movement and Production in a Local Environment) model of evolution of cooperation allow for the emergence of social dilemmas.  In the SIMPLE model, benefits diffuse to nearby patches in each time period, leading productive agents to have positive effects on the fitness of nearby individuals.  When agents have the ability to follow a positive gradient towards benefits, productive individuals 'attract' neighbors, leading to changes in density and assortment.  This leads to complex feedback loops between evolutionary and spatial dynamics that can favor the evolution and maintenance of cooperation.  Agents are embedded in a one-dimensional ring of patches that moves down time as time progresses, providing a rich visual record of agent movement and benefit production over the course of the simulation.  The agent-based nature of this model and the accessibility of information about agent behavior over time in the visual record can help to cultivate intuitions about complex spatial and evolutionary processes including those underlying multilevel selection, tradeoffs between present and future resource availability, and the impact of environmental variability/uncertainty on optimal strategies.


-- 

C. Athena Aktipis
Department of Psychology
University of Pennsylvania

AktipisSWARMFEST-abstract.doc
Description: MS-Word document