Agent Modeling, Behavioral Complexity, and Wicked Problems: Lessons Learned from a Multi-Disciplinary Collaborative Agent-Based Simulation of Water Management in a Transboundary Region

How much behavioral complexity can be integrated into real-world agent-based modeling, and does it matter whether we capture this complexity when considering wicked environmental problems? Early social science advocates of agent-based modeling insisted that understanding the interactions between actors through networks would shed light on the emergent qualities of behavior that could, in turn, provide the richest understanding of human complexity. We describe the collaboration that developed between a team of cultural, psychological, and environmental anthropologists and a team of socio-environmental modelers in order to show how we were able to balance human behavioral complexity with the quantitative and computational constraints of creating an agent-based simulation of water management. At the same time, we describe how a traditional reliance on network-based approaches failed to meet the needs of our team. Our collaboration focused on the U.S./Mexico Rio Grande/Rio Bravo and was complicated by the issue of wicked problems that challenged traditional system dynamics models: multiple political and legal boundaries, cultural difference, different approaches toward land and resource management and conservation, and different historical traditions. We summarize our new simulation model for the region (implemented in ENVISION) and focus our discussion on the value of — and pitfalls associated with — creating working water manager typologies out of significant socio-cultural and individual diversity and complexity.