Navigating unsafe behavioral decision-making in workplaces: An agent-based simulation for intervention and management

Background Workers’ unsafe behaviors primarily cause accidents in workplaces. Despite extensive research on the antecedents of unsafe behaviors, understanding workers’ dynamic decision-making under the influence of safety cognition, group interactions, management, site conditions, and behavioral outcomes as a holistic system remains limited. Objective This study aims to explore workers’ safety performance across different site risk environments and various intervention measures, including the effects of intervention frequency, punishment strictness, and interaction intensity on dynamic safety performance in low, medium, and high-risk environments. Methods An agent-based simulation model considering multiple factors of individual safety cognition and behavioral decision-making was conducted. It examined the complex roles of workers’ cognitive decision-making processes within workgroup interactions, management strategies, and environmental contexts. Results The results reveal that: (a) management interventions effectively reduce unsafe behaviors across all risk environments; (b) punishment strictness significantly inhibits unsafe behaviors, with moderate punishment being most effective in low-risk environments; (c) the impact of worker interactions varies significantly by risk level—limiting interactions in low-risk environments, promoting interactions in medium-risk environments, and additional measures are needed in high-risk environments to reduce unsafe behaviors. Conclusion This research lays a new foundation for the causes of workers’ unsafe behaviors when corresponding with effective management strategies. It also provides valuable insights for project managers in cognitive management and policy analysis.