Introduction

The management of emergency activities such as guiding people out of dangerous areas and coordinatingrescue teams is characterized by uncertainty regarding both the source of danger and the availability ofuseful resources. Depending upon the scale and nature of the incident, people involved in a crisis maysuffer from limited situational awareness (SA). SA involves being aware of what is happening around inorder to understand how information, events, and the crowd actions will impact the goals and objectives…

Conclusions

In this paper, we presented an early-stage development results on OpenStack cloud-based multiagent simulation platform for evacuation of the crowd from the indoor environment with the limited number of evacuation exits and evacuation path size. Environment in this model is represented by cell automata and interpreted as a potential field, in which generated agents are located. The crowd management in the cloud is supported by the MapReduce programing model with the classical Hadoop framework used for its implementation. Simple experiments were performed on a small Hadoop cluster with ten nodes and separately for a single powerful server in order to demonstrate potential benefits of using the cloud system. The results of the experiments show that cloud-based systems can reduce significantly the complexity of the management of individuals in the crowd. Moreover, there is no need to initiate the large number of new processes on the same work station cause some data processing operations can be performed by using the software frameworks shared inside the public cloud…