Agent-Based Simulation to Predict Occupants’ Physical-Distancing Behaviors in Educational Buildings

With a long-term global pandemic impact, physical-distancing is recommended as the most efficient strategy for defending individuals, by lowering the risk of community spreading, especially for indoor spaces. It is important to understand how environmental constraints and individual physical-distancing behaviors affect total risks of physical distance violation at the collective level, so as to better design and implement facility operational measures to reduce COVID transmission risks.

Researchers developed an agent-based simulation model to test the physical-distancing policies in an educational facility based on multiple layers of occupancy behaviors that can occur.

Modeling Physical-Distancing Behaviors

Using the pedestrian simulation capabilities of AnyLogic, researchers designed and conducted simulation experiments to evaluate the impact of measures to secure physical distancing. In particular, the impact of three scenarios was examined:

  • limiting classroom occupancy;
  • adjusting breaktime;
  • different social norms on physical-distancing.

Agents in the model represent occupants in the education facility (e.g., students, staff), and their default behaviors are governed by rules that include movements between destinations (e.g., classrooms, restrooms, waiting areas, gates) with the consideration of predetermined schedules (e.g., class durations, breaktime duration). They basically follow the shortest route to their destinations and will avoid possible collisions with other objects (e.g., other agents, physical structures, obstacles).

Effects of Behavior Securing Physical Distance
Effects of Behavior Securing Physical Distance ((a) congestions at the start of breaktime; (b) accumulated late agents at the start of class time)

An agent’s internal physical-distancing intent will affect its behavior in two ways:

  • in determining a specific arrival location within its destination space, the agent will select the arrival location considering the physical distances to other agents to be temporarily fixed within the destination space;
  • when the agent is moving toward its destination, its internal intent will function as a social force to lower its movement acceleration based on the distances with other agents with longer adjusting time.


The results show the impacts of capacity, break time duration, and internal intent to secure physical distancing on the risks of violating physical distancing.

The results indicate that spatial and temporal constraints of educational environments could greatly diminish the impact of preventive operational measures for physical distancing, by creating various bottlenecks. This finding indicates that such simulation of physical distancing behaviors could benefit educational administrators and facility managers in designing optimal operational measures when considering their educational facilities and student populations.

Related posts