A Hybrid Mesoscopic/Agent-Based Model for Crowd Dynamics with Emotional Contagion

This work investigates crowd dynamics in evacuation scenarios by incorporating the effects of emotional contagion. We develop a hybrid modeling framework that couples a mesoscopic kinetic approach for pedestrian motion with an agent-based model describing emotional interactions. Each individual adapts their emotional state based on the average emotion of nearby agents, which in turn affects their walking direction and speed. Numerical simulations, conducted using a Monte Carlo particle method for the kinetic motion model together with an agent-based scheme for emotional contagion, examine three representative scenarios: calm crowds without contagion, fully emotional crowds, and mixed populations. The results show that emotional contagion intensifies congestion and significantly increases evacuation times. The proposed dual-scale model provides a first step toward a comprehensive representation of panic dynamics, offering new insights into the interplay between motion and emotion, with potential applications to crisis management and safety design.