Open Agent-Based Modeling Consortium

Description:
Intellectual Merits
The dynamic complexity that characterizes interactions between humans and the natural environment, has intersected over the past century with increasingly rapid population growth, urbanization, and technological development to make human society an important driver of environmental changes that threaten to exceed our abilities to adapt using traditional means. This makes it imperative that we find better ways to track these global socio-ecological systems (or ‘socioecosystems'), and anticipate their social and natural consequences if we are to lessen the risk of increasingly severe socio-natural catastrophes like those that shook the world in the past year. Late 20th Century advances in information technology offer powerful new tools to assist us in understanding-and hopefully even managing-these coupled social and natural systems. In this context, Agent Based Models (ABMs) have recently emerged as a promising cybertool to study the dynamics of complex human and biological systems, integrating individual perceptions and behaviors in the contexts behavioral ecology, game theory of decision-making, and geospatial representations of the world. While ABMs are much discussed and are rapidly becoming perceived as a requirement for cutting-edge research centering on human environmental interaction, most social and natural scientists still have a limited awareness of their potential, and the experimental nature of most ABM platforms prevents them from being readily accessible to researchers. Furthermore, a lack of standards has inhibited the exchange of modeling expertise and cumulative model building among different research teams, and there has been little effort at systematically addressing problems of validation and verification in modeling algorithms and results.

We propose to address these issues and accelerate the regular integration of ABMs for research in social-natural dynamics by initiating a broad, community-wide agenda involving active researchers ranging from content experts in the social sciences, ecology, conservation biology and GIS, to computer scientists and mathematicians. To accomplish this, we will organize a workshop of leading practitioners to establish a Community Modeling Framework for Socioecological research (CMFS), following successful examples of community frameworks for cybertool development in other research domains. We, as organizers-a geoarchaeologist, ecologist, and computer scientist-represent the diversity of fields we propose to articulate, and the workshop participants span the range scientists employing ABM in socio-ecological research and ABM platform developers. The workshop will be followed by a proof-of-concept pilot project, comprising a new collaborative, scientific network to initiate the CMFS. The pilot project will establish a modeling archive and component library for jump starting new research using ABM, a collaboration server (CVS) for improving usability and usefulness of ABM for socioecological research, a testbed of standard data for developing model evaluation protocols, proposed best practices for model dissemination and frameworks for model interchange, and a training program in ABM aimed at social and natural scientists. The Center for Social Dynamics and Complexity (Arizona State University) and Santa Fe Institute will co-sponsor this endeavor.

Broader Impacts
The workshop will represent the first forum directed at establishing a community-wide framework to promote ABM use for socioecological research. It will encompass a broad diversity of researchers, including minorities, individuals with international and multicultural experience, and early career scientists. Students will be directly involved with organizing the workshop and developing the CMFS, and along with professionals, will be targeted with planned training in ABM applications. If existing community frameworks for cyberinfrastructure and related research serve as a guide, the community modeling framework we propose to establish has the potential for far-reaching impacts on the way we investigate the complex interactions of humans and the environment-through promoting a much wider expertise in cybertools directly aimed at studying complex systems. Because questions about the dynamics of socio-ecological systems lie at the heart of increasingly critical policy debates at global scale, the long-term agenda of the program we are initiating in this proposal has the potential to influence future environmental policy the way climatic modeling does today.