To Friday 15 September 2023
By Unit Penjenamaan, FSKTM
By employing an agent-based model, researchers can analyze crowd behavior and improve evacuation procedures in educational environments.
Safety during emergencies is of paramount importance, this is particularly true especially in densely populated educational environments.
Understanding and predicting individual behaviour during such instances can greatly increase the effectiveness of evacuation procedures. Traditionally, live drills and physical experiments have been the go-to methods for this purpose.
However, these practices can be time-consuming, expensive, and incapable of capturing the full breadth of possible human behaviours.
At the Faculty of Computer Science and Information Technology (FCSIT) at UNIMAS, researchers led by Hamizan Sharbini are turning to agent-based models—a type of computer simulation—as a solution.
These simulations mirror the behaviour of individuals within a crowd, functioning like a 2D particle agents where each agent can react independently and make its own decisions.
In the face of an emergency, such as a fire alarm ringing, swift and safe evacuation of all occupants is vital. In these simulations, the agents represent the students who must understand the evacuation protocol for such emergencies during the learning process.
The agent-based model considers a variety of factors that can influence an individual’s behaviour during evacuation, such as walking speed, desired exit point, reactions of nearby individuals, and potential obstacles.
By running the simulation, researchers can observe the overall behaviour of the crowd without putting real humans at risk during mock drills.
From these simulations, valuable insights can be drawn, such as potential bottlenecks where crowding may occur or ways to optimize evacuation routes.
This data is invaluable for emergency planners, architects, and responders, assisting in the design of buildings and planning of evacuation procedures.
In essence, an agent-based model provides an innovative tool for studying and understanding crowd behaviour during emergencies. By effectively replicating a real-world emergency drill, it helps inform strategies that can enhance safety and ensure a swift and incident-free evacuation.
The application of agent-based models in understanding crowd behaviour is a significant step towards creating safer educational environments.
By studying the simulation data, researchers hope to provide insights that can reduce the risks associated with emergency evacuations in academic institutions.
This research is funded by Universiti Malaysia Sarawak under PILOT Research Grant Scheme.