Modeling and simulation of stranded passengers' transferring decision-making on the basis of herd behavior

According to the case of large crowd being stranded at transportation hubs (such as bus station, railway station, airport, etc.) in climate disasters, a model is designed to represent passengers' transferring among transportation hubs on the basis of herd behavior. Bayesian network and influence diagram are used in model to capture essential features of the passenger's decision-making process, then considering the influence of group behavior on the individuals, the transition probabilities of passengers are revised with passengers' accuracy of belief and group information. A multi-agent simulation model is developed in Repast platform in Java language. Through the simulation data, the influence of herd behavior over stranded passengers' transferring mode and system performance in different policies are analyzed, some policy recommendations for the government are proposed.