Agent Architecture for Evacuation Pedestrian’s Crowded Public Places: Combination of Individual and Collaborative Reasoning

A major problem in a crowd evacuation is rescuing frightened and trapped people. This paper presents a decision-making process based on a combination of individual and social collaborative reasoning to evacuate people during a crowd evacuation. Specifically, we propose to evacuate crowds by modeling the social behaviors of multi-agent systems. In this context, collaborative reasoning contributes to finding data on emergency exits. Collaborative reasoning allows agents to exchange information and reach an agreement to exit quickly and safely. Experimental results showed that collaborative reasoning combined with individual reasoning has a great impact on the final decision and can lead to a successful evacuation with a percentage of 90.