Modeling and Optimization of Crowd Guidance for Building Emergency Evacuation

Abstract

Effective building evacuation in case of emergencies such as fires, chemical spills, or spreads of biological agents has long been recognized as an important issue, and effective crowd guidance can improve egress efficiency, occupant survivability, and mitigate or prevent undesirable consequences such as blocking or stampeding. To effectively guide crowds, however, is a challenging issue because emergency events may propagate in uncertain ways and affect the availability of egress paths; egress path capacities may constrain the speed of crowd movement; and crowd could be stressed making their behaviors different from their normal modes. Although good results have recently been obtained on microscopic behaviors of individuals such as the social force model of Helbing [1-3], there is a major gap between these microscopic models of individuals evacuating from rooms and the macroscopic models of crowd flows needed to evacuate them from a building. Most existing egress guidance methods assume that crowd behaviors are independent of emergency situations and are fully controllable under guidance. These assumptions make it difficult to capture important features such as stampeding or blocking.