crowdsimulation

Crowdsimulation refers to the computational modeling of the movement and interactions of large groups of people within a shared space. It is used to analyze safety, optimize design, plan for events, and create believable crowd behavior in virtual environments.

Most approaches are microscale, simulating individual agents with goals and perception, though meso- and macro-level models treat the crowd as a continuous density. Prominent micro-models include the social force model, cellular automata floor-field models, and agent-based simulations that encode local decision rules.

Key components include the environment geometry, obstacles, exits, and agent properties such as preferred speed, personal space, and route preferences. Agents perceive local surroundings, plan routes, and adjust motion to avoid collisions and congestion.

Applications span evacuation planning, venue and corridor design, crowd management at events, public-safety simulations, urban-scale mobility studies, and the generation of realistic crowd scenes in films and video games.

Calibration and validation rely on empirical trajectory data from video analysis, sensor networks, or controlled experiments. Limitations include model simplifications, variability in human behavior, scalability challenges, and ethical considerations related to privacy and surveillance. Simulations inform risk assessment but should be interpreted with uncertainty.