Modeling study of pedestrian evacuation considering dynamic guidance under terrorist attack events

Abstract

An extended cellular automaton model incorporating the interaction among attackers, dynamic guidance, and evacuees is proposed to study pedestrian evacuation behavior under the threat of terrorist attack. In this model, the assault behaviors of attackers described by physical self-driving forces are involved. Multiple fields including the static floor field, group driving field, attack threat field and following-guide field are introduced to study movement rules of evacuees and guiders. Especially, how multi-source information, namely optimal exit route and attack threat information are disseminated by guides and absorbed by evacuees are further investigated. Simulation results show that there is an optimal number of guiders to minimize the crowd evacuation time and casualties, and the optimal number of guiders is significantly correlated with exits number rather than attackers. The initial position configuration of guiders aiming to improve evacuation effectiveness needs to be determined according to the pedestrian density, it is beneficial to arrange guiders near exits at high pedestrian density and in the middle of room at low-density situations. An increase in the information propagation radius of guiders is conducive to evacuation time and crowd survival rates to some extent. Pedestrian evacuation time becomes longer and crowd fatalities are more severe if attackers are more inclined to assault evacuees compared to guiders.