Modeling Human Behavior During Fire in Feed Mills: A Hybrid Agent-Based and Discrete Event Simulation Approach

This paper presents an innovative model that integrates hybrid agent-based and discrete event simulation approaches to model human behavior during fire incidents in animal feed factories. The model takes into account both physical factors—such as building structure, escape routes, stairwells, and exit doors—and the distinct behavioral responses and decision-making patterns of employees and visitors during an emergency, thereby creating a realistic evacuation scenario. Simulation results demonstrate that trained personnel can manage evacuation processes more effectively, whereas untrained visitors negatively impact the process due to slower reaction times and a lack of situational awareness. Notably, the combination of dual stairwells and dual exits was found to minimize bottlenecks by allowing individuals to disperse in different directions, significantly reducing evacuation times. The hybrid model further incorporates individual decision-making mechanisms, panic responses, and communication networks into the simulation, offering a dynamic and multidimensional analysis that extends beyond traditional models. The findings underscore the critical importance of addressing both human behavior and structural factors in fire safety planning, while also highlighting the need for additional precautions in taller and more complex structures. Overall, the developed model serves as a valuable tool for formulating strategies aimed at optimizing evacuation processes during fire emergencies.