A review of cellular automata models for crowd evacuation

With the increasing of risk potential in crowded places, evacuation management becomes practically important to ensure the safety of crowds. The studies of crowd evacuation in normal or emergency situations have become a hot topic. Due to the distinct advantages of high efficiency, strong scalability and simple implementation, cellular automata models (CA) have become one of the most widely-used models for evacuation. However, the practical requirements of evacuation propose some important challenges for CA models, for example, to accurately characterize both position and velocity of individuals, to depict environments and accidents, and to describe human behaviors. In the last 20 years, there are many studies aiming at resolving the above challenges. Starting from the challenges mentioned above, this paper tries to give a review of CA models, specially used for crowd evacuation. Firstly, we give an overview of CA models for evacuation, and put forward research paradigm, modeling framework and classification of CA models. The models used for evacuation are classified into three kinds of categories, i.e. lattice gas model, floor field model, and other field-based models. The last category includes potential field model, electrostatic-induced potential field model, cost potential field model, etc. Then, three main challenges of CA models for evacuation are presented, and the improvements for each type of challenge are summarized. Typical simulation scenarios and research issues are further proposed. Finally, the advantages and disadvantages of CA models are illustrated from the aspects of implementation, performance, scalability, accuracy and applicability.