An Agent-Based Model to Investigate Different Behaviours in a Crowd Simulation

This paper presents an agent-based model to evaluate the effects of different behaviours in a crowd simulation. Two different behaviours of agents were considered: collaborative, acting attentively and collaboratively, and defector who, on the other hand, acts individually and recklessly. Many experimental simulations on different complexity scenarios were performed and each outcome indicates how the presence of a percentage of defector agents helps and motivates the collaborative ones to be better and more fruitful. This investigation was carried out considering the (i) number of agents evacuated, (ii) exit times and (iii) path costs as evaluation metrics.