The result, as visualized in Fig. 3, illustrates the annual number of published PES models that used the four different PES approaches. The figure suggests that models based on the CA and ABM approaches account for the majority of existing PES models, and that the ABM approach has gradually become more prevalent over time.
This figure shows the evacuation simulation results based on the CIM platform to demonstrate the three-dimensional dynamic multi-layer evacuation process.
This figure shows the spatial layout plan formed by the evacuation simulation software and visualized crowd evacuation routes. The node coordinates are firstly extracted from the layout drawing of the building carried by CAD and then imported into the software. At the initial data setting stage, a specific layout drawing of the building can be designed for given a scenario. In addition, the location of fires and parameters such as numbers of evacuating agents can be customized. During simulation, the evacuation route will be displayed in real time on the map.
This figure shows the controller of the crowd evacuation simulation software, namely FREEgress (Fire Risk Emulated Environment for Egress), which was independently developed by our team. The software can generate new simulations or replay previous simulation results as required. Displayed items can be customized according to users’ needs. Currently, there are six options available: map, evacuees' vision, group network, evacuation reference point coordinates, floor objects, and evacuation trajectory.
This figure shows the evacuation path-finding mechanism of agents. The behavior model of agents consists of the following three parts: perception, decision-making and execution. First, the agent updates the perceived information of environment and crowd through its sensors. Then in the decision-making stage, the agent selects corresponding behaviors according to specific rules based on the perceived information and its own state, and finally in the execution stage the agent executes the selected behaviors.
An agent-based simulator for indoor crowd evacuation considering fire impacts (FREEgress demo)
This study presents the collaboration work between the research group of Prof. Kincho Law, working in the Department of Civil and Environmental Engineering, Stanford University and Dr. Nan Li, working in the Department of Construction Management, Tsinghua University. In this work, a new fire evacuation simulation model, named FREEgress (Fire Risk Emulated Environment for Egress), is developed to simulate the dynamic influences of heat, temperature, toxic gas and smoke particles on evacuees’ mobility, navigation decision making and health conditions, showing in our video demo.
Details can refer to:
Li, Z., Huang, H., Li, N., Chu, M. L., & Law, K. (2020). An agent-based simulator for indoor crowd evacuation considering fire impacts. Automation in Construction, 120, 103395.
Studying indoor crowd emergent evacuation is of great significance for ensuring human safety and reducing casualties. With the development of computer science, simulation of crowd evacuation in fire scenarios has become the main direction of current evacuation research. The purpose of this research is to build evacuation simulation models based on multi-agent modeling method that considers real-time impact of fire factors, in order to achieve more accurate simulations of crowd evacuation in building fire scenarios, and then conduct efficient analysis and visualization of the evacuation process in the City Information Modeling (CIM) platform based on the simulation results. Our research results can help analyze and understand the crowd evacuation patterns in fire scenarios at the level of academic research, and benefit the fire management department to predict and manage fire emergency evacuation process at the level of application. It also contributes to evaluation of evacuation performance and optimization of performance-based fire protection design of buildings.
Jieyu Chen, Xiaolu Xia, Ziwei Li, Haifeng Yang, Huang Huang
the National Natural Science Foundation of China (NSFC) under Grant No. 71603145, 2017.01-2019.12
the Humanities and Social Sciences Fund of the Ministry of Education of China (MOE) under Grant No. 16YJC630052，2015.01-2017.12
the Tsinghua University-Glodon Joint Research Centre for Building Information Model (RCBIM)
