Enhancing evacuation safety in urban primary schools: an agent-based model integrating child development behaviour and health dynamics

Primary schools require specific evacuation plans considering children's developmental vulnerabilities, but current models frequently overlook children's behavioural and health dynamics. Our proposed reliability-driven agent-based modelling (ABM) framework incorporates the Theory of Planned Behaviour (TPB) for decision-making and Dose-Response Model (DRM) for quantifying health hazards associated with smoke exposure, aiming to close this gap. Using real-world evacuation data from an urban primary school, we simulate risk scenarios to assess interactions between evacuation efficiency, environmental risks, and safety regulations. Results show that smoke density significantly affects system reliability, increasing evacuation time by 43.94% (±0.834) and decreasing evacuee health by 76.83% (±0.735) in the worst-case scenario. We found that more adult guides during evacuations could compensate for students' unpreparedness. However, teacher placement's efficacy and interaction vary greatly depending on smoke patterns. This outcome highlights the need for contingency-based evacuation plans customised to different environmental circumstances and acknowledges the challenges in evaluating the effectiveness of school evacuation strategies. This work contributes to schools’ system safety by offering adaptive solutions for dynamic risks, emphasising the importance of child-specific reliability specifications in school emergency planning. The framework offers practical ideas for urban resilience, enabling organisations to mitigate health hazards and enhance the safety of critical infrastructures in uncertain environments.