Multi-Agent Simulation Collision Avoidance of Complex System: Application to Evacuation Crowd Behavior

In this work, we present a collision avoidance technique for a crowd robust navigation of individuals in evacuation which is a good example of a complex system. The proposed algorithm is inspired from the Reynolds model, with the addition of several individuals' behavioral criteria as well as a microscopic perception of the environment, which affects their travel speeds and emerging appeared phenomena. Our system is modeled by agent and tested by a Netlogo simulation, several modules such as A* planning, physical and psychological factors of agents have been programmed and successfully inserted into a 3D environment. Our application can be used as a framework to simulate real situations (evacuation of a stadium, a building...) in order to arrive at strategies to decision support of a complex system, which is a real problem in our daily life.