Crowd Dynamics Simulation Research

摘要

This paper reports on research aimed at providing arnmore realistic simulation model of crowd behaviorrnunder both panic and non-panic conditions and both inrna constrained (room) and unconstrained (outside)rnenvironment by considering individual differences inrnage, gender, culture and ethnicity. Such simulationsrnwill be of great value both in operational and trainingrncontexts. Many operations deal with crowd control andrnaccurate simulations of crowd dynamics. The explicitrnincorporation of gender, ethnicity, age, and culturalrndifferences as factors in the model will ensure broadrnapplicability of the research to crowd control in nationsrnother than the U.S.rnThe approach is based on a Social Force Model whichrnconsists of two components: social interaction andrnphysical interaction. The social interaction relies on anrnindividual’s desire for personal space and the physicalrninteraction consists of body compression and friction.rnTo be generalizable beyond modeling of crowds in thernU. S., the simulation needs to capture the effect ofrnculture on personal space. Culture definitely plays arnkey role in the size of personal space. For example,rnJapanese prefer a larger surrounding space thanrnAmericans, while Italians prefer a smaller personalrnspace. Physical contact between cultural groups alsornvaries. Additionally, to be a widely accepted, usablernmodel, it will be necessary to validate the behavior ofrnthe entities in the simulation against the behavior ofrnreal crowds captured on video. This validation willrnalso help to fine tune the model parameters to reflectrndifferent crowd compositions.rnThese results will support research in sociology,rneconomics, and psychology. They will also providernimproved implementation of emergency managementrnsimulations thereby providing broad societal benefits.rnAdditionally architectural design for work and publicrnspaces or for egress during panic conditions will bernimproved.