Yiyu WANG, University of Leeds , United Kingdom
Jiaqi GE, University of Leeds , United Kingdom
Alexis COMBER, University of Leeds , United Kingdom
Research on effective crowd evacuation has become increasingly relevant due to the growing number of public places with large-scale crowd gatherings. Owing to the uncertainty of pedestrian-related evacuation factors (e.g. response time, emotions, social networks, bounded rational evacuation strategies), various obstacles (e.g. lack of comprehensive data and understanding of behavior) need be overcome to support in-depth studies [1,2]. Therefore, developing an effective simulation model of pedestrian flow with a true reflection of individual behaviors can be beneficial to establish large-scale crowd evacuation planning.
Since agent-based modeling and simulation has become one of the mainstream approaches for pedestrian research, this research developed an evacuation simulation model of pedestrian flow based on Muti-agent system (MAS) and Bayesian Nash equilibrium (BNE). After comparison, BNE was adopted to overcome the propensity for a relatively uniform distribution of the pedestrians in evacuation space, and it shortens evacuation time by avoiding massive congestions . MAS was used to simulate their movement and behaviors. Several rules (e.g. pedestrian speed-density relation, collision avoidance rules, etc.) underpinning the model for movement aspect was defined to improve simulation accuracy. Some theories (e.g. following/conformity behavior, etc.) were introduced to complement the rationality of pedestrian decision-making process during evacuation. This simulation model was constructed in Netlogo. According to initial running results, the following behavior linked to BNE utility is evident in model’s evacuation patterns. Further research direction is to adopt other pedestrian behaviors (e.g. competitive behavior) and social factors (e.g. emotions) to improve the simulation accuracy and efficiency.
Keywords: Agent-based modeling|pedestrian behaviors|Bayesian Nash equilibrium|emergency evacuation|pedestrian movement