COMPUTATIONAL SOCIAL SCIENCE

CSS SEMINAR - ROGERS

Friday, November 21, 3:00 p.m.
Center for Social Complexity Suite
Research Hall, Third Floor

Archaeological Use of Agent-Based Models: Some Challenges and Emerging Directions

J. Daniel Rogers, Ph.D.
Department of Anthropology
National Museum of Natural History
Smithsonian Institution

and

Wendy Cegielski, Ph.D. Student
School of Human Evolution and Social Change
Arizona State University

ABSTRACT: For the last 100 years archaeologists have investigated cultural differences and the dynamics of long-term change. These studies have relied on interpretation of material remains and their depositional context as the sources of information. By the 1960s, new paradigms had developed that focused on process over description and this movement ushered in an era of multivariate statistical analyses and inferences. Although the first computational simulations were run in the 1970s, disciplinary counter-currents emphasized alternative narrative perspectives effectively discouraging further exploration into contributions that computational simulations might make to the field. After 2000, however, a substantial growth in the use of agent-based models can be documented. In this presentation we evaluate examples of published results that illustrate different computational modeling strategies currently employed in archaeology. In a time of dramatic growth it is appropriate to make a critical assessment of the successes along with persistent challenges and emerging opportunities. We have identified several larger issues: the perception of simulations as overly complex, the lack of clarity in the appropriate use of simulations, and the integration of simulation into established fields. As we move forward with this research we are actively seeking input about what practitioners see as the major challenges and emerging directions in computational social science.