A multiagent-based model for pedestrian simulation in subway stations

Multiple components such as complicated station environment, heterogeneous pedestrians and diverse interactions make it difficult to realize an effective pedestrian simulation in subway stations. A multiagent-based model is proposed on the basis of the metamodel proposed by Béhé, which includes the subway station environment abstraction model, three-level pedestrian agent model and interactive rule base. The first term incorporates some notions to satisfy the specific modeling demands of subway stations and uses the object-oriented modeling method to realize a normative, extensible and flexible simulation environment building platform. The second integrates the three levels of NOMAD model and takes staged pedestrian behaviors into account, which works well for simulating real pedestrian behaviors in subway stations. The third realizes various and staged interactions correctly and automatically. Finally, a case study of Ping'anli Station in Beijing Subway is presented and three comparison issues are introduced to validate the effectiveness of our models.