基于系统动力学提出了建立兼顾模拟精度和模拟效率的地下空间群体疏散模型的新思路. 借助专业软件Vensim进行定量模拟, 在总初始人数为分别为350、400、450、500、 1000、2000、\4000 七种工况下, 模拟结果与专业模拟软件Pathfinder的模拟结果相近, 且计算速度更快. 该模型可再现``快即是慢''等群体疏散行为中的典型现象. 通过对大规模群体疏散过程中的关键影响因素``总初始人数'' ``恐慌状态''进行分析, 发现了以下现象: 1) 总初始人数过多会导致系统疏散能力降低, 疏散结束时间大幅滞后; 2) 恐慌程度过低会导致系统疏散效率长时间保持在较低水平, 不利于被困人群的及时疏散. }

The modeling research on large-scale pedestrian evacuation flow in underground space has been a hot topic in recent years. In this study we have developed a new model based on system dynamics theory in order to analyze the phenomena and other correlative questions have formed the theories mechanism of pedestrian characteristic of the large-scale evacuation in underground space. The result of this system dynamics model is very close to the Pathfinder which has been widely used for the commercial analysis. For seven different conditions (the overall number of people is 350/400/450/500/1000/2000/4000), the system dynamics model value is less than 7\% and the $R^2$ is 0.9997. This system dynamics model can be applied to simulate the large-scale evacuation without limitations on population, and also the well-known phenomenon, ``faster-is-slower'', is easily reproduced via this model. The simulation was carried out based on two major factors: ``overall population'' and ``panic mood'', and the impact of the evacuation capacity of export was studied. It was observed that large overall population result in evacuation time delay and lower level of panic mood lead to inefficient evacuation.