Integration of Emotional-Cognitive Architecture in ACT-R for Multi-Agent Social Dynamics for Disaster Response Simulation

Accurately simulating human behavior in disaster scenarios is crucial for effective emergency management and social response prediction. However, existing approaches lack integration of emotional factors in cognitive modeling, leading to limited fidelity in behavioral simulation. This study presents an enhanced ACT-R framework incorporating an emotional module to simulate post-disaster social dynamics, with a specific focus on earthquake scenarios. We propose a novel integration methodology that combines ACT-R with Agent-Based Modeling (ABM), where emotional states are dynamically modeled through a triple-parameter system (emotional intensity, social support level, and resource availability) that adaptively influences cognitive processing and decision thresholds. By implementing this emotion-aware framework in a case study of 20,000 agents within Xi'an city wall area, we observed distinct emotional clustering patterns and collective safety-seeking behaviors driven by emotional contagion, demonstrating the framework's capability in capturing complex social dynamics during crisis scenarios.