A Preliminary Study of Individual Based Crowd Simulation Based on Bayesian Nash Equilibrium

The lack of experimental datasets for individual behaviours has hindered the systematic studies of pedestrian behaviours as well as the refined development of regular laws of individual movement in simulation models. This research developed a simulation model for crowd evacuation on the basis of Bayesian Nash Equilibrium (BNE) and a Multi-Agent System (MAS). BNE was introduced in this research to augment the rationality of individual decision-making process in evacuation simulation and to assist pedestrians in discovering an optimal evacuation route to avoid congestions. A series of simulation experiments were conducted to evaluate the performance of the initial model, and the current experimental results demonstrate a noticeable positive influence of BNE on reducing evacuation time. A detailed introduction of the establishment and implementation details of the model as well as model analysis have been provided in this paper. Limitations and a few optional research directions in the future are also discussed.