An Interdisciplinary Approach to Agent-based Modeling of Conflict in Eastern Africa

! An Interdisciplinary Approach to Agent-Based Modeling of Conflict in Eastern Africa By Ian Skoggard and William G. Kennedy Introduction Over the past four years, a team of computational social scientists at the Center for Social Complexity at George Mason University (GMU) in collaboration with anthropologists from the Human Relations Area Files (HRAF) at Yale University have been developing a multi-scale spatial agent-based model to better understand the environmental, social, and cultural dimensions of conflicts in the Rift Valley region of eastern Africa. The overall goal of the joint GMU-HRAF project is “to build and analyze innovative and interrelated computational models of asymmetric conflict with explicit sociocultural content that can advance understanding and improve policy analysis.” The anthropologists’ contribution is to provide the “explicit sociocultural content” for the models. In this paper, we discuss the interdisciplinary collaboration in developing the agents for two models, a prototype model called “HerderLand” concerned with pastoral movement in a savanna environment and “RiftLand,” a 1,600 sq km region in eastern Africa. The paper focuses on the give and take between the anthropologists and computational social scientists in developing the agents. The challenge for the anthropologists is how to apply their knowledge, both ethnographic and theoretical, in shaping a model that ultimately will help policymakers anticipate and manage conflict in eastern Africa. Modeling and Anthropology Modeling is not new to applied anthropology. Indeed, any construing of cultural norms and values can be regarded as a model of human behavior, usually a verbal model (Zacharias, MacMillan, and Van Hemel 2008). However, in the 1950s, system theorists began applying