Paper “Social Distancing and Behavior Modeling with Agent-Based Simulation” is presented at the CAAD Future 2021 conference and inclusion in the CAAD Futures 2021 Springer Proceedings.
Presentation. 16 – 18 JULY 2021.
The research discusses applying agent-based simulation (ABS) technology to analyze the social distancing in public space during the COVID-19 pandemic to facilitate design and planning decisions. The ABS is used to simulate pedestrian flow and construct the micro-level complexity within a simulated environment. This paper describes the various computational methods related to the ABS and design space under the new social distancing guidelines. We focus on the linear phases of agent activities, including (1) environmental query, (2) waiting in a zone, (3) waiting in a queue, and (4) tasks (E-Z-Q-T) in response to design iterations related to crowd control and safety distance. The design project is extended to the agents’ interactions driven by a set of tasks in a simulated grocery store, restaurant, and public restroom. We applied a quantitative analysis method and proximity analysis to evaluate architectural layouts and crowd control strategies. We discussed social distancing, pedestrian flow efficiency, public accessibility, and ways of reducing congestion through the intervention of the E-Z-Q-T phases.
Keywords: agent-based simulation, social distancing, crowd control
Figure 3. Agent density and space proximity map. ABS without social distancing vs. with social distancing rules. Each agent’s autonomous “action” lies in modifying its movement based on its rules and environment. Top. Floor plan and interior perspective of a check-in area of a restaurant. Middle: proximity map without social distancing. Bottom: proximity map with 2-meter social distancing with the same number of agents in the same given time. Notice the hot waiting areas’ issues are replaced with a larger waiting area, while some agents choose not to walk in the restaurant after EQ. Right. Compare the number of occupancies. Red: agents with social distancing. Blue: agents without social distancing.
This research was funded by UC Forward, as a part of the Price Hill project at UC.