A Review of Building Evacuation Models - NIST Virtual Library
A Review of Building Evacuation Models - NIST Virtual Library
A Review of Building Evacuation Models - NIST Virtual Library
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
A.13 Legion<br />
Developer: Legion International, Ltd., UK<br />
Purpose <strong>of</strong> the model: The purpose <strong>of</strong> this model is to aid in space planning and optimization<br />
through the prediction <strong>of</strong> crowd behavior as an interaction between individuals 66-68 . The model<br />
can be used for a wide variety <strong>of</strong> applications (i.e. railway and metro stations, airports, and tall<br />
buildings) and needs (i.e. design, refurbishment, and operation and safety assessment).<br />
Availability to the public for use: This model has been commercially available through Legion<br />
International Ltd. since May 2003.<br />
Modeling method: This is a behavioral model.<br />
Structure <strong>of</strong> model: The Legion model works in a vector 2D continuous space, instead <strong>of</strong><br />
superimposing a coarse or fine grid network onto the floor plan. In addition, by providing a<br />
continuous approach to the structure configuration, the model can simulate counterflow,<br />
overtaking, people and obstacle avoidance, and negotiation through crowds. The model refers to<br />
its structure as an “unbounded choice” method. This method explores the possible moves<br />
available to the occupant in vector space which is updated constantly, instead <strong>of</strong> being<br />
constrained by a set <strong>of</strong> rules.<br />
Perspective <strong>of</strong> model and occupant: The model views the occupants as individuals. Each<br />
individual in the model is considered to be a virtual person and is simulated accordingly with<br />
distinct physical and psychological characteristics and objectives.<br />
The occupant’s view <strong>of</strong> the building is also an individual perspective. This virtual person moves<br />
in a realistic manner. Occupants determine their path based on their perception and information<br />
stored in the space.<br />
Occupant behavior: Artificial Intelligence. Legion views the occupants as intelligent<br />
individuals and social, physical, and behavioral characteristics are assigned probabilistically<br />
from empirically established pr<strong>of</strong>iles. The social characteristics include gender, age, culture, and<br />
pedestrian type (i.e. commuter versus tourists) which Legion states shape typical movement<br />
preferences. The physical characteristics addressed are body size. And, the behavioral<br />
characteristics include memory, willingness to adapt, and preferences for unimpeded walking<br />
speeds, personal space, and acceleration. These characteristics make up a pr<strong>of</strong>ile for each person<br />
and are based on distributions derived from video footage <strong>of</strong> actual pedestrians.<br />
Also, Legion allows for conditional and probabilistic behavior to be superimposed on the base<br />
model.<br />
Occupant movement: Occupant movement within the model is in agreement with extensive<br />
empirical research performed on the study <strong>of</strong> crowd movement and behavior. Research teams<br />
have acquired and analyzed video footage <strong>of</strong> individual and crowd behavior. Movement is based<br />
upon the “least-effort principle,” which means that each individual attempts to minimize<br />
dissatisfaction (defined as the aggregate <strong>of</strong> frustration, inconvenience, and discomfort). These<br />
A-48