Simulation of Cognitive Pedestrian Agents Crowds in Crisis Situations

Margaret Lyell, Mark Becker

In crisis situations in an urban environment, first responder teams often must deal with crowds of people. Consider the case of a building fire in a dense city environment. People may be injured; walkways may be blocked, with fire equipment attempting to reach the scene. Crowd behavior can become an issue when trying to reach the injured, ensure safety and restore conditions to normal. The motivations of pedestrians that form the crowd can vary. Some are there because they are curious about the crisis situation. Others, attending to their individual concerns, may have found themselves in the ‘wrong’ location. They may be trying to leave the area, but the density of people as well as the spatial layout of the walkways may be impeding their progress. Other individuals, unaware of the fire, may be attempting to reach their intended destinations that happen to be near the crisis area, thus adding to crowd congestion. With a model of crowd behavior, effective strategies for resource usage in managing crowd behavior can be developed. Our approach to this problem is that of agent-based modeling and simulation. We develop a cognitive pedestrian agent model. Utilizing this model, we simulate crowd behavior in a ‘city fire’ scenario. Characteristics of crowd behavior with different pedestrian personality mixes and a strategy for crowd management are investigated

Full Text