Agent-based climate governance model integrates cognitive and institutional dynamics

A research team led by Ivan Puga-Gonzalez has proposed a multi-level agent-based architecture that unifies cognitive and institutional dynamics for climate governance simulations. Presented at the AAMAS 2026 workshop on Agents for Societal Impact, the 9-page paper details an approach that goes beyond existing models by integrating empirically grounded decision-making with strategic institutional behavior. The framework combines three key components: (i) individual-level decision-making using the HUMAT (human motive-based) and MOA (motivation-opportunity-ability) frameworks, (ii) social influence propagation through demographic homophily networks, and (iii) institutional strategy modules representing environmental NGOs, media agents, and politicians.

The architecture is notable for its emergent political decision-making, which aggregates multiple signals including expert advice, public mobilization levels, party alignment, and media framing. This enables realistic simulation of democratic climate governance processes. The model is designed for empirical calibration using synthetic populations derived from survey data and institutional parameters from Living Lab stakeholder engagement. Rather than presenting results, this paper focuses on design principles, modular structure, and integration logic, outlining pathways for future validation and generalization to other policy domains. The approach offers a more holistic way to study how individual behaviors and institutional strategies interact to shape climate policy outcomes.