An evacuation simulation model of pedestrian flow using Bayesian Nash equilibrium and a Multi-Agent System

Abstract

Abstract. The shortage of experimental data for individual behaviours has hampered systematic research on pedestrian behaviours and the refined development of evidence informed rules for pedestrian movement in simulation models. This research proposes an evacuation simulation model of pedestrian flow based on Bayesian Nash Equilibrium (BNE) and Multi-Agent System (MAS). In this paper, BNE was used to augment the logics of pedestrian decision-making processes in an evacuation MAS simulation and to improve evacuating agents’ movement and behaviours. A detailed introduction of the construction process and implementation details for the initial model as well as the visualization of experiment results are provided in this paper. Limitations and several potential future research directions are also identified.