Emergency crowd evacuation modeling and simulation framework with cellular discrete event systems

Abstract

First responders and security personnel face many challenges to safely evacuate crowded environments. Conducting frequent physical and table-top exercises are neither feasible nor economical. This is where modeling and simulation comes into play in providing a risk-free and economical method to practice various evacuation strategies, train first responders, and provide accurate decision-support and emergency guidance. With the aid of formal methods, here we present a suite of various egress strategies build on top of an open-source modeling and simulation environment. We use Cellular Discrete-Event System Specification (Cell-DEVS) formalism to explore emergency evacuation scenarios by building 12 egress models representing aspects of human behaviors under emergencies and the activities of authorities in guiding the crowd. The proposed framework explores random and controlled human movement, as well as implementing psychological conditions such as herd following and panicked states. We provide results analysis to compare evacuation speed under various egress methods, allowing for decision making when using the suite for training purposes. The outcome of this work is available on a public repository to serve the DEVS community, researchers, and public safety authorities interested in emergency evacuation simulations.