Agent-based modeling of emergency evacuation in a railway station square under sarin terrorist attack

Abstract

Evacuation modeling is a promising measure to support decision making in scenarios such as flooding, explosion, terrorist attack and other emergency incidents. Given the special attention to the terrorist attack, we build up an agent-based evacuation model in a railway station square under sarin terrorist attack to analyze such incident. Sarin dispersion process is described by Gaussian puff model. Due to sarin’s special properties of being colorless and odorless, we focus more on the modeling of agents’ perceiving and reasoning process and use a Belief, Desire, Intention (BDI) architecture to solve the problem. Another contribution of our work is that we put forward a path planning algorithm which not only take distance but also comfort and threat factors into consideration. A series of simulation experiments demonstrate the ability of the proposed model and examine some crucial factors in sarin terrorist attack evacuation. Though far from perfect, the proposed model could serve to support decision making.