APPLICATION OF AGENT-BASED SIMULATION IN EMERGENCY EVACUATION AND CASUALTY PREDICTION OF INDOOR SCENARIOS

Simulation of crowd emergency evacuation is a complex part in which the objectives are improving the safety of the crowd by assessing geometrical structure of the escape area to decrease evacuation time and optimize architectural design and rescue operations. Usually, in public infrastructures, casualties can increase by inappropriate design of the architectural elements and intensify when hazardous high density crowd movements arise. In these cases, evacuation encounters a problem which can lead to injury or death of people. It is obvious that evacuation behaviors of the occupants will affect the casualty number in earthquakes. Furthermore, the structural failure patterns during an earthquake can influence occupants casualties. Besides, investigating how people behave in the event of a disaster (like earthquake) can lead to a more accurate prediction of the occupants casualty of the building. Microscopic modeling approaches are recognized as a valuable basis to study occupant dynamics under emergency conditions. In this paper, evacuation models that mostly used or presented in last years such as cellular automata, social force, and agent-based models are presented. Besides, the studies related to the agent-based modeling that mostly used in indoor scenarios are reviewed. Then, Netlogo software presented as a tool for agent-based modeling and used for the simulation of a case study. Results have been analyzed by the Netlogo software under different failure scenarios of the building caused by an earthquake excitation. The case study is a classroom of a one-story high school building in China that studied by Li et al. (2018). Simulating the behavior of occupants need adequate information that can be obtained from a real-life video recording. For verification of the model, the data of the video recording such as speed and path finding patterns are applied. The evacuation process is shown in Figure 1 at different time intervals. Dark circles present students evacuate from the classroom and red circle is a random selected student for showing each position during time instants.

Keywords:

Evacuation , Agent-based modeling , Netlogo , Earthquake , Casualty